login fix maybe

app.py

@@ -258,6 +258,19 @@ with gr.Blocks(title="Moderation Model Testing") as demo:
         ]
     )
     
+    # Update token status on app load (when OAuth token is already available)
+    demo.load(
+        handle_login_click,
+        inputs=None,  # OAuth token auto-injected
+        outputs=[
+            token_status_markdown,
+            save_mode_help,
+            refresh_private_btn,
+            refresh_roost_btn,
+            dataset_help_text,
+        ]
+    )
+    
     # Dataset load handler
     def load_example_from_dataset(selected_label, cached_examples_list, dropdown_choices_list):
         """Load example from dataset and populate all fields."""
@@ -327,6 +340,19 @@ with gr.Blocks(title="Moderation Model Testing") as demo:
         ],
     )
 
+    # Update token status on app load (when OAuth token is already available)
+    demo.load(
+        handle_login_click,
+        inputs=None,  # OAuth token auto-injected
+        outputs=[
+            token_status_markdown,
+            save_mode_help,
+            refresh_private_btn,
+            refresh_roost_btn,
+            dataset_help_text,
+        ]
+    )
+
 
 if __name__ == "__main__":
     demo.launch(ssr_mode=False)

ui/sidebar.py

@@ -18,11 +18,13 @@ def build_sidebar() -> dict:
             """
             This interface allows you to test moderation models with custom content policies.
             
-            **🧪 Testing Tab**: Enter content to test against your policy. View model predictions, categories, reasoning traces, and raw responses.
+            **🧪 Testing Tab**: Enter content to test against your policy. View model predictions, categories, reasoning traces, and raw responses. Optionally save results to datasets.
             
             **📋 Policy Definition Tab**: Define your content policy by uploading a markdown file, entering it manually, or selecting from preset examples.
             
             **⚙️ Configuration Tab**: Select models, adjust generation parameters, and customize system prompts and response formats.
+            
+            **📊 Session Management Tab**: Browse and load previously saved test results. Load examples from your personal dataset or the shared ROOST dataset.
             """
         )