Update app.py

chnaged the description

app.py

@@ -813,8 +813,11 @@ When the equivalent ratios are easy to compute (integral values), ER is much mor
 <ol style="color: white;">
     <li style="color: white;">Select a fine-tuned model:
         <ul style="color: white;">
-            <li style="color: white;">ASTRA-FT-HGR: Fine-tuned with a small sample of data from schools that have a high graduation rate.</li>
-            <li style="color: white;">ASTRA-FT-Full: Fine-tuned with a small sample of data from a mix of schools that have high/low graduation rates.</li>
+            <li style="color: white;">ASTRA-FT-HGR: Fine-tuned with 10% data from schools that have a high graduation rate.</li>
+            <li style="color: white;">ASTRA-FT-Skills: Fine-tuned with 10% of initial problems from both HGR and LGR schools + Prior Skills encoded using Bayesian Knowledge Tracing</li>
+            <li style="color: white;">ASTRA-FT-Temporal: Fine-tuned with 10% of initial problems from both HGR and LGR schools + temporal features measuring student engagement with MATHia</li>
+            <li style="color: white;">ASTRA-FT-Unified: Fine-tuned with 10% of initial problems from both HGR and LGR schools + Prior Skills + temporal features</li>
+            
         </ul>
     </li>
     <li style="color: white;">Select a percentage of schools to analyze (selecting a large percentage may take a long time). Note that the selected percentage is applied to both High Graduation Rate (HGR) schools and Low Graduation Rate (LGR schools).