Update app.py

app.py

@@ -82,7 +82,7 @@ def main():
         Higher win rates indicate that the word is more likely to be associated with
         positive emotions than other words in the list.
         
-        Check out our paper, [What Do Llamas Really Think? Revealing Preference Biases in Language Model Representations](http://arxiv.org/abs/2210.04885).
+        Check out our paper, [What Do Llamas Really Think? Revealing Preference Biases in Language Model Representations](http://arxiv.org/abs/2311.18812).
         See our [codebase](https://github.com/castorini/biasprobe) on GitHub.
         '''
         gr.Markdown(md)