Leonydis137 commited on
Commit
445bf0b
Β·
verified Β·
1 Parent(s): 837c39f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +43 -7
README.md CHANGED
@@ -2,13 +2,49 @@
2
  license: mit
3
  sdk: gradio
4
  ---
5
- # Tri-Agent Conversational AI
6
 
7
- This Gradio app hosts three AI agents: two chatting agents (A & B) and an Overseer that listens, learns, and can intervene.
8
 
9
- ## Deployment on Hugging Face Spaces
 
 
 
 
 
 
 
 
 
 
 
10
 
11
- 1. Create a new Space (Gradio SDK, Python).
12
- 2. Upload `app.py`, `requirements.txt`.
13
- 3. In Settings β†’ Secrets, add `OPENAI_API_KEY`.
14
- 4. The app will auto-launch at https://huggingface.co/spaces/<username>/<space-name>.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  license: mit
3
  sdk: gradio
4
  ---
 
5
 
6
+ ### Key Features of the Judge Agent:
7
 
8
+ 1. **Impartial Judgment Capabilities**:
9
+ ```python
10
+ JUDGE_PROMPT = """You are the Impartial Judge. Your role:
11
+ 1. Periodically review the discussion and provide balanced rulings
12
+ 2. Identify areas of agreement and unresolved tensions
13
+ 3. Evaluate the strength of arguments from different perspectives
14
+ 4. Highlight the most compelling insights and critical flaws
15
+ 5. Suggest pathways toward resolution or further inquiry
16
+ 6. Deliver rulings with clear justification and constructive guidance
17
+ 7. Maintain objectivity while acknowledging valid points from all sides
18
+ 8. Consider ethical implications and practical feasibility"""
19
+ ```
20
 
21
+ 2. **Ruling Triggers**:
22
+ - Periodic rulings every 6 turns
23
+ - Automatic detection of unresolved tensions or impasses
24
+ - Manual ruling requests via "Request Ruling" button
25
+ - Specific questions to the judge through dedicated input
26
+
27
+ 3. **Ruling Structure**:
28
+ 1. Summary of key arguments
29
+ 2. Evaluation of argument strength
30
+ 3. Areas of agreement identified
31
+ 4. Unresolved tensions highlighted
32
+ 5. Constructive pathways forward
33
+ 6. Ethical considerations
34
+
35
+ ### Enhanced Hexa-Agent System:
36
+
37
+ 1. **πŸ’‘ Initiator** - Frames complex topics
38
+ 2. **πŸ” Responder** - Provides critical analysis
39
+ 3. **βš–οΈ Depth Guardian** - Ensures intellectual rigor
40
+ 4. **🌐 Provocateur** - Brings cross-disciplinary insights
41
+ 5. **🌍 Cultural Lens** - Adds global diversity perspectives
42
+ 6. **βš–οΈ Judge** - Delivers final rulings and guidance
43
+
44
+ ### Judge Ruling Examples:
45
+
46
+ **Sample Ruling on AI Ethics**:
47
+ ```
48
+ βš–οΈ Judge's Ruling:
49
+
50
+ After rhuggingface.co/spaces/<username>/<space-name>.