update limit and markdown

app.py

@@ -190,7 +190,7 @@ class FiftyOneAssistant:
         """Wrapper for GPU embedding function."""
         return get_embedding_gpu(text, config.embedding_model)
     
-    def hybrid_search(self, query_text: str, query_vector: List[float], limit: int = 2, text_boost: float = 0.3):
+    def hybrid_search(self, query_text: str, query_vector: List[float], limit: int = 4, text_boost: float = 0.1):
         """
         Hybrid search using only dense vectors + text filtering
         """
@@ -689,6 +689,8 @@ def create_demo():
         # 🚀 VoxelGPT-Mini: A FiftyOne Assistant
 
         Your **intelligent code companion** for computer vision and machine learning workflows using FiftyOne!
+
+        NOTE: This assistant can make mistakes! It's using a small 1B parameter model and retrieving from a synthetic dataset created from FiftyOne documentation, tutorials, and other content.
         
         ## ✨ Powered By:
         - **LLM Engine**: Meta's Llama 3.2 (1B) with fine-tuned adaptation for FiftyOne expertise