Update description

app.py

@@ -53,15 +53,15 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("# VLMVibeEval")
     gr.Markdown(
         """
-        A lightweight leaderboard for evaluating Vision Language Models (VLMs) — based on vibes.
+        A lightweight leaderboard for evaluating Vision Language Models (VLMs) — based on vibes. 🌞
 
-        Traditional benchmarks can be misleading due to overlap with training data. Instead, we let you **vibe test** models across curated examples:
+        Traditional benchmarks don't give concrete signal for your use case and models are often saturated over them. Instead, we let you **vibe test** models across curated, in-the-wild examples:
         
         1. Predefined categories with images and prompts.
         2. Check any model on these examples.
-        3. Explore the generations and judge for yourself.
+        3. Explore the generations and judge for yourself, as different models have different styles and strengths. 🗣️
 
-        This is not about scores — it's about *how it feels*.
+        This is not about scores — it's about *how it feels*. You can submit new models in the community tab and we'll shortly update the app! 🤗
         """
     )