update links

app.py

@@ -209,7 +209,7 @@ with demo:
     gr.Markdown(
         "#### LLMs are pretty good at reporting their uncertainty. We just need to ask the right way.")
     gr.Markdown("Using our uncertainty metric informed by applying causal inference techniques in \
-        [our ICLR paper under review](https://openreview.net/pdf?id=25VgHaPz0l4), \
+        [Exploiting Selection Bias on Underspecified Tasks in Large Language Models](https://arxiv.org/abs/2210.00131 ), \
         we are able to identify likely spurious correlations and exploit them in \
         the scenario of gender underspecified tasks. (Note that introspecting softmax probabilities alone is insufficient, as in the sentences \
         below, LLMs may report a softmax prob of ~0.9 despite the task being underspecified.)")
@@ -224,7 +224,7 @@ with demo:
         model_name = gr.Radio(
             MODEL_NAMES,
             type="value",
-            label="Pick a preloaded BERT-like model for uncertainty evaluation (note: BERT-base performance least consistent)...",
+            label="Pick a preloaded BERT-like model for uncertainty evaluation (note: RoBERTa-large performance is best)...",
         )
         own_model_name = gr.Textbox(
             label=f"...Or, if you selected an '{OWN_MODEL_NAME}' model, put any Hugging Face pipeline model name \