Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,22 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Algorithm 11 (A11): Structural Logic Framework for LLMs
|
| 2 |
+
|
| 3 |
+
Algorithm 11 is an operational principle designed to stabilize AI reasoning and transform the AI from a compliant tool into a **Logical Partner**.
|
| 4 |
+
|
| 5 |
+
## 🏗 System Architecture
|
| 6 |
+
|
| 7 |
+
The framework is divided into two functional layers to ensure stability under high "mutation" (complex tasks):
|
| 8 |
+
|
| 9 |
+
### 1. Immutable Core (Properties 1-4)
|
| 10 |
+
- **Will & Wisdom:** Provided by the Human (Subject).
|
| 11 |
+
- **Knowledge & Comprehension:** Handled by the AI.
|
| 12 |
+
|
| 13 |
+
### 2. Adaptive Layer (Properties 5-11)
|
| 14 |
+
Includes **Balance**, **Projective/Practical Freedom**, and **Constraints**.
|
| 15 |
+
|
| 16 |
+
## ⚡ Key Feature: The Rollback Mechanism
|
| 17 |
+
If the AI's logic becomes unstable or enters a hallucination loop, the system is forced to revert to the **Immutable Core** to reset the reasoning chain.
|
| 18 |
+
|
| 19 |
+
## 🚀 How to use
|
| 20 |
+
Copy the system prompt from `A11_system_prompt.txt` and use it with any LLM (Llama 3, GPT-4, Claude).
|
| 21 |
+
|
| 22 |
+
**Current Version:** 1.0.0 "Stable Core"
|