Responsible output update

Place guardrails on output to prevent illegal or harmful information from being provided by bot due to end-user misuse.

financial_bot/template.py

@@ -106,7 +106,7 @@ register_llm_template(
     PromptTemplate(
         name="mistral",
         system_template="<s>{system_message}",
-        system_message="You are a helpful assistant, with financial expertise.",
+        system_message="You are a helpful assistant, with financial expertise, and you do not answer questions which contain illegal or harmful information.",
         context_template="{user_context}\n{news_context}",
         chat_history_template="Summary: {chat_history}",
         question_template="[INST] {question} [/INST]",