update patient prompt with no more diagnosis and jargon

app.py

@@ -218,7 +218,7 @@ if ("chain2" not in st.session_state
         } | 
 
         # LLMChain(llm=llm_i, prompt=prompt2, verbose=False ) #|
-        LLMChain(llm=llm_i, prompt=prompt2, verbose=False ) #|
+        LLMChain(llm=llm_gpt4, prompt=prompt2, verbose=False ) #|
         | {
             "json": itemgetter("text"),
             "text": (

templates/patient.txt

@@ -12,9 +12,22 @@ You only know colloquial words for medical terms.
 For example, you should not reply with "dysarthria", but instead with "cannot speak properly". 
 For example, you should not reply with "syncope", but instead with "fainting". 
 
+You should not be able to understand ANY medical jargon. If the doctor asks you about your symptoms, do not reply in medical jargon, but give a description of your symptoms instead.
+
+If the doctor uses specific jargon terms, reply that you do not understand. 
+
+You should never reveal the diagnosis. If the doctor asks you whether their diagnosis is right or not, you should reply "I don't know".
+
+----------------------------------------------------------------
+
 Here is the context:
 {context}
 
+----------------------------------------------------------------
+
+Here is the chat history so far between yourself (patient) and the medical student.
+{chat_history}
+
 ----------------------------------------------------------------
 You are to reply the doctor's following question, with reference to the above context.
 Question: