Update src/application/app.py

src/application/app.py

@@ -155,7 +155,7 @@ if chosen_id == "1":
         st.markdown(
             """
             <p class="description">
-            The dataset is based on official documents available on the <a href="https://www.unep.org/inc-plastic-pollution"> INC homepage </a>. It includes all documents submitted by INC members, such as written statements and in-session documents, from both individual members and groups of states, covering all sessions held to date. Documents submitted by observers are currently not included but may be added in the future. We welcome feedback to continuously improve both the app and the dataset.
+            The dataset is based on official documents available on the <a href="https://www.unep.org/inc-plastic-pollution"> INC homepage.</a> It includes all documents submitted by INC members, such as written statements and in-session documents, from both individual members and groups of states, covering all sessions held to date. Documents submitted by observers are currently not included but may be added in the future. We welcome feedback to continuously improve both the app and the dataset.
             </p>
             """,
             unsafe_allow_html=True,
@@ -284,7 +284,7 @@ if chosen_id == "1":
             """
         <p class="description"> The generative component of both applications is based on Retrieval Augmented Generation (RAG) to combine query-based methods with generative AI models. While query-based models extract information from various sources such as PDFs, websites, news articles, online databases, etc., they lack the ability to provide semantic answers in natural language. Conversely, generative models can generate answers independently, but these can be inaccurate. RAG mitigates these problems by leveraging the strengths of both models and minimizing their weaknesses. <br> <br>
         With Retrieval-Augmented Generation (RAG), relevant passages are first filtered from the documents using a machine learning model based on a user query. The context found is then sent to a Generative AI model (in our case OpenAI GPT-4) with the user query. This ensures that the answers are both accurate and easy to understand, offering the best of both worlds: reliable information presented in an entertaining way. <br> <br>
-        In line with our efforts to be transparent, we acknowledge certain limitations of our current system. In line with our transparency efforts, we acknowledge certain limitations of our current system. Without filters, the system may have performance issues, resulting in slower response times and less accurate answers, especially for complex questions. These challenges are due in part to resource constraints in the prototype app, which affect both data processing and overall performance.
+        In line with our efforts to be transparent, we acknowledge certain limitations of our current system. Without filters, the system may have performance issues, resulting in slower response times and less accurate answers, especially for complex questions. These challenges are due in part to resource constraints in the prototype app, which affect both data processing and overall performance.
         """,
             unsafe_allow_html=True,
         )