Update app.py

app.py

@@ -82,14 +82,14 @@ with gr.Blocks() as demo:
   # Phi-2 Chatbot Demo
   This chatbot was created using Microsoft's 2.7 billion parameter [phi-2](https://huggingface.co/microsoft/phi-2) Transformer model. 
   
-  In order to reduce the response time on this hardware, `max_new_tokens` has been set to `21` in the text generation pipeline. With this default configuration, it takes approximately `60 seconds` for the response to start being generated, and streamed one word at a time. Use the slider below to increase or decrease the length of the generated text.
+  In order to reduce the response time on this hardware, `max_new_tokens` has been set to `128` in the text generation pipeline. With this default configuration, it takes approximately `60 seconds` for the response to start being generated, and streamed one word at a time. Use the slider below to increase or decrease the length of the generated text.
   """
     )
 
     tokens_slider = gr.Slider(
         8,
         128,
-        value=21,
+        value=128,
         label="Maximum new tokens",
         info="A larger `max_new_tokens` parameter value gives you longer text responses but at the cost of a slower response time.",
     )