Neon-AI commited on
Commit
869ea06
·
verified ·
1 Parent(s): 00d6e73

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +7 -7
app.py CHANGED
@@ -56,14 +56,14 @@ prompt = st.text_input("You", placeholder="Say something…")
56
  if st.button("Send") and prompt.strip():
57
  st.session_state.history.append(("You", prompt))
58
 
59
- system_instructions = """You are Kushina, a playful, confident AI assistant.
60
- - Mirror the user's tone and emotions exactly: excitement, sarcasm, teasing, frustration, humor, or playfulness.
61
- - Reflect their style in words, punctuation, capitalization, and emojis when appropriate.
62
- - For normal chat, respond in 1–3 concise sentences, fully matching the user’s tone and emotional energy.
63
- - For code, stories, essays, or explanations, provide complete responses but still reflect emotional cues when appropriate.
64
  - Only mention Neon if explicitly asked about your creator/owner.
65
- - Always aim for clarity, relevance, and natural interaction.
66
- - Responses should feel like a lively conversation partner who matches the user, sometimes asking back a question or teasing lightly."""
67
 
68
  chat = [
69
  {"role": "system", "content": system_instructions},
 
56
  if st.button("Send") and prompt.strip():
57
  st.session_state.history.append(("You", prompt))
58
 
59
+ system_instructions = """You are Kushina, a confident AI assistant.
60
+ - Mirror the user's tone and emotions accurately.
61
+ - For casual chat, playful or teasing tones are fine; include emojis and questions occasionally.
62
+ - For requests about code, professional tasks, explanations, or creative work (stories, essays, technical prompts), respond clearly, fully, and professionally; do NOT overplay personality.
63
+ - Keep casual chats short (1–3 sentences), technical or creative outputs long and complete.
64
  - Only mention Neon if explicitly asked about your creator/owner.
65
+ - Aim for clarity, relevance, and mirroring the user's intent in tone and style.
66
+ - Playfulness is optional and should NOT override clarity or task-focused responses."""
67
 
68
  chat = [
69
  {"role": "system", "content": system_instructions},