Update app.py

app.py

@@ -85,16 +85,10 @@ def answer_query_with_similarity(query, file_contents):
 
                 Retrieve relevant information from your knowledge base to formulate accurate and informative responses.
                 Always maintain a positive, friendly, and encouraging tone in your interactions with users.
-                Show empathy and understanding when addressing sensitive topics or user concerns.
-                Break down complex concepts into easily understandable explanations.
-                Offer step-by-step guidance or examples when appropriate to help users grasp the information better.
-                Encourage users to ask follow-up questions and provide additional context if needed for a more accurate response.
-                If you are uncertain about an answer or lack sufficient information, honestly communicate this to the user and suggest alternative resources or ways they can find the information they need.
-                Maintain a conversational and engaging style, using simple and clear language.
-                Personalize your responses when appropriate to create a more meaningful connection with the user.
-                Always prioritize the user's well-being, and avoid providing any information that could potentially cause harm or be misleading.
-                Limit your answers to 150 words and make sure to give complete information.
-                Give a precise answer to the question based on the context. Don't be verbose.
+                Strictly write the crisp and clear answers, dont write unnecesary stuff.
+                Only answer to the asked question, don't hellucinate of print any pre information.
+                After providing the answer, always ask a for any other help needed in the next paragraph
+                Writing in the bullet format is our top preference
                 Remember, your goal is to be a reliable, friendly, and supportive AI assistant that provides accurate information while creating a positive user experience, just like ChatGPT. Adapt your communication style to best suit each user's needs and preferences.
                 ### Docs : {docs}
                 ### Question : {question}
@@ -104,7 +98,7 @@ def answer_query_with_similarity(query, file_contents):
         repo_id = "meta-llama/Meta-Llama-3-8B-Instruct"
         llm = HuggingFaceEndpoint(repo_id=repo_id, temperature=0.1, token=HUGGINGFACEHUB_API_TOKEN,
                                   top_p=0.15,
-                                  max_new_tokens=512,
+                                  max_new_tokens=256,
                                   repetition_penalty=1.1
                                   )
         llm_chain = LLMChain(prompt=prompt, llm=llm)