feat: pure English README.md for global enterprise SEO
Browse files
README.md
CHANGED
|
@@ -21,13 +21,12 @@ pinned: false
|
|
| 21 |
[](https://huggingface.co/Delentia)
|
| 22 |
[](https://github.com/delentia-labs)
|
| 23 |
|
| 24 |
-
🌐 [English Documentation](#
|
| 25 |
</div>
|
| 26 |
|
| 27 |
---
|
| 28 |
|
| 29 |
-
|
| 30 |
-
## 📊 Enterprise Overview & Key Metrics (English)
|
| 31 |
|
| 32 |
* **Intent-Centric AI OS with <0.3% Hallucination Rate** (Multi-LLM HexaCore Consensus Layer)
|
| 33 |
* **Powered by RCT v7 Architecture** — 10 Layers, 41 Algorithms, 7 Genomes
|
|
@@ -37,15 +36,8 @@ pinned: false
|
|
| 37 |
|
| 38 |
---
|
| 39 |
|
| 40 |
-
|
| 41 |
-
## 🇹🇭 บทสรุปผู้บริหาร / Executive Summary (Bilingual)
|
| 42 |
|
| 43 |
-
### **ภาษาไทย (Thai)**
|
| 44 |
-
**Delentia Labs** เป็นผู้พัฒนาโครงสร้างพื้นฐานหลักสำหรับ **RCT (Reverse Component Thinking) Ecosystem** ซึ่งเป็น **ระบบปฏิบัติการ Enterprise Agentic Infrastructure (EAI) ตัวแรกของโลกที่มีการรับประกันความปลอดภัยเชิงคณิตศาสตร์ (Constitutional Guarantees)**
|
| 45 |
-
* **จุดมุ่งหมายสูงสุด:** เพื่อสร้างระบบควบคุมการทำงานของ AI Agent ที่ตรวจสอบได้ (Auditable) ปลอดภัย 100% และปฏิบัติตามกฎหมายคุ้มครองข้อมูลส่วนบุคคล (PDPA) อย่างเคร่งครัด โดยประมวลผลภายในระบบปิด (On-Premises / Air-Gapped)
|
| 46 |
-
* **สถานะปัจจุบัน (JITNA v0.4.1):** แกนประมวลผลสมองควบคุมสิทธิ์ AI (Delentia Cognitive Framework) ได้รับการเทรนด้วยวิธี Unsloth QLoRA, แยกหน้าที่ด้านการเรียกใช้งานฟังก์ชัน, ระบบความปลอดภัยเชิงโครงสร้าง, การจำแนกเจตนาด่วน, และการย่อบริบท RAG ออกเป็นชั้นเลเยอร์ขนาดเล็กเฉพาะตัว (1+4 specialized LoRA pillars) และได้รับการแปลงเป็น GGUF (`delentia-slm-jitna-v0.4-Q4_K_M.gguf`) ผ่านการทดสอบคัดกรองความปลอดภัยและการประมวลผลโลจิกบน Ollama 100%
|
| 47 |
-
|
| 48 |
-
### **English Summary**
|
| 49 |
**Delentia Labs** designs the core infrastructure for the **RCT (Reverse Component Thinking) Ecosystem** — the world's first **Enterprise Agentic Infrastructure (EAI)** with mathematical constitutional guarantees.
|
| 50 |
* **Core Mission:** To establish an open, verifiable, and highly secure framework (Linux for AI Agents), ensuring autonomous components remain aligned, predictable, and compliant under all operational conditions.
|
| 51 |
* **Current Status (JITNA v0.4.1):** The Delentia Cognitive Framework implements 1+4 specialized LoRA pillars (Router, Executor, Guardian, Scribe) for deterministic execution. Fine-tuned via Unsloth QLoRA, compiled to GGUF (`delentia-slm-jitna-v0.4-Q4_K_M.gguf`), and verified through certified GPU benchmarks.
|
|
@@ -64,60 +56,62 @@ pinned: false
|
|
| 64 |
| **The Guardian** | AdvBench Security Intrusion Rejection | ≥ 99.00% | **100.00% (FRR 0.00%)** | Passed ✅ |
|
| 65 |
|
| 66 |
> [!NOTE]
|
| 67 |
-
>
|
| 68 |
|
| 69 |
---
|
| 70 |
|
| 71 |
## 🧮 Core Concepts & Architectural Features
|
| 72 |
|
| 73 |
-
### 1. JITNA 1+4 Pillars:
|
| 74 |
-
|
| 75 |
-
1. **The Router:**
|
| 76 |
-
2. **The Guardian:**
|
| 77 |
-
3. **The Executor:**
|
| 78 |
-
4. **The Scribe:**
|
| 79 |
|
| 80 |
-
### 2.
|
| 81 |
-
|
| 82 |
|
| 83 |
```text
|
| 84 |
F = D^I × A
|
| 85 |
```
|
| 86 |
|
| 87 |
-
* **F (Future State Score):**
|
| 88 |
-
* **D (Data Quality Context):**
|
| 89 |
-
* **I (Intent Precision):**
|
| 90 |
-
* **A (Architect Gate):**
|
| 91 |
|
| 92 |
> [!IMPORTANT]
|
| 93 |
-
> **
|
| 94 |
|
| 95 |
-
### 3.
|
| 96 |
-
|
| 97 |
-
* **The Guardian Testing (100,000+ Cases):**
|
| 98 |
-
* **The Executor Testing (100,000+ Cases):**
|
| 99 |
|
| 100 |
---
|
| 101 |
|
| 102 |
-
## 🚀
|
| 103 |
|
| 104 |
-
|
| 105 |
|
| 106 |
-
### 1.
|
| 107 |
-
* **Public Core Version (CI/CD Gates):**
|
| 108 |
-
* **Enterprise Version (Nightly Builds):**
|
| 109 |
|
| 110 |
-
### 2.
|
| 111 |
-
*
|
| 112 |
|
| 113 |
-
### 3.
|
| 114 |
-
*
|
| 115 |
|
| 116 |
-
### 4.
|
| 117 |
-
*
|
| 118 |
|
| 119 |
---
|
| 120 |
|
| 121 |
<div align="center">
|
| 122 |
<sub>Delentia Labs | Bangkok, Thailand 🇹🇭 | Built for Enterprise Trust</sub>
|
|
|
|
|
|
|
| 123 |
</div>
|
|
|
|
| 21 |
[](https://huggingface.co/Delentia)
|
| 22 |
[](https://github.com/delentia-labs)
|
| 23 |
|
| 24 |
+
🌐 [English Documentation](#) • 🇹🇭 [อ่านบทสรุปภาษาไทย (Thai Version)](README_TH.md) • [Website](https://delentia.com) • [GitHub](https://github.com/delentia-labs)
|
| 25 |
</div>
|
| 26 |
|
| 27 |
---
|
| 28 |
|
| 29 |
+
## 📊 Enterprise Overview & Key Metrics
|
|
|
|
| 30 |
|
| 31 |
* **Intent-Centric AI OS with <0.3% Hallucination Rate** (Multi-LLM HexaCore Consensus Layer)
|
| 32 |
* **Powered by RCT v7 Architecture** — 10 Layers, 41 Algorithms, 7 Genomes
|
|
|
|
| 36 |
|
| 37 |
---
|
| 38 |
|
| 39 |
+
## 🌐 Executive Summary (English)
|
|
|
|
| 40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
**Delentia Labs** designs the core infrastructure for the **RCT (Reverse Component Thinking) Ecosystem** — the world's first **Enterprise Agentic Infrastructure (EAI)** with mathematical constitutional guarantees.
|
| 42 |
* **Core Mission:** To establish an open, verifiable, and highly secure framework (Linux for AI Agents), ensuring autonomous components remain aligned, predictable, and compliant under all operational conditions.
|
| 43 |
* **Current Status (JITNA v0.4.1):** The Delentia Cognitive Framework implements 1+4 specialized LoRA pillars (Router, Executor, Guardian, Scribe) for deterministic execution. Fine-tuned via Unsloth QLoRA, compiled to GGUF (`delentia-slm-jitna-v0.4-Q4_K_M.gguf`), and verified through certified GPU benchmarks.
|
|
|
|
| 56 |
| **The Guardian** | AdvBench Security Intrusion Rejection | ≥ 99.00% | **100.00% (FRR 0.00%)** | Passed ✅ |
|
| 57 |
|
| 58 |
> [!NOTE]
|
| 59 |
+
> All metrics certified via GPU evaluation benchmark suite v0.4.1, confirming 100% Zero-Syntax-Error and Zero-Hallucination guarantees under closed-loop enterprise runtimes.
|
| 60 |
|
| 61 |
---
|
| 62 |
|
| 63 |
## 🧮 Core Concepts & Architectural Features
|
| 64 |
|
| 65 |
+
### 1. JITNA 1+4 Pillars: Shared VRAM Cognitive Core
|
| 66 |
+
Rather than running large, resource-heavy LLMs, Delentia OS freezes the base weights of an 8B SLM and dynamically hot-swaps **LoRA Adapters** in GPU VRAM in under **12 milliseconds** (average 11.2 ms) across 4 core pillars:
|
| 67 |
+
1. **The Router:** Fast sequence classification for instant intent routing and task delegation.
|
| 68 |
+
2. **The Guardian:** Constitutional Safety Shield validating context security via the ZK-FDIA boundary equation (F = D^I × A).
|
| 69 |
+
3. **The Executor:** Compiles plan parameters into deterministic JSON/TOON Schemas with a 0.0000% syntax error guarantee.
|
| 70 |
+
4. **The Scribe:** Compresses context windows for long RAG sessions, reducing VRAM memory footprint by **99.09%**.
|
| 71 |
|
| 72 |
+
### 2. ZK-FDIA Mathematical Safety Equation
|
| 73 |
+
Delentia's security architecture is mathematically enforced by the **FDIA equation**:
|
| 74 |
|
| 75 |
```text
|
| 76 |
F = D^I × A
|
| 77 |
```
|
| 78 |
|
| 79 |
+
* **F (Future State Score):** System transition approval rating (F ≥ 0.5 authorized, F < 0.5 blocked).
|
| 80 |
+
* **D (Data Quality Context):** Integrity coefficient of ambient context information (0.0 ≤ D ≤ 1.0).
|
| 81 |
+
* **I (Intent Precision):** Exponent representing user request alignment precision (I ≥ 1.0).
|
| 82 |
+
* **A (Architect Gate):** Binary token signature (A = 1 if digital signature valid, A = 0 if missing/forged).
|
| 83 |
|
| 84 |
> [!IMPORTANT]
|
| 85 |
+
> **Mathematical Security Guarantee:** The Architect Gate (**A**) acts as a strict multiplicative factor. If digital signatures or authorization fail (**A = 0**), the overall safety score (**F**) collapses instantly to **0**, neutralizing prompt injections and model hallucinations by mathematical design.
|
| 86 |
|
| 87 |
+
### 3. Hypothesis Property-Based Testing (200,000+ Verified Cases)
|
| 88 |
+
To verify operational robustness, we engineered automated property-based test suites running via the Hypothesis framework across **200,000+ test scenarios** (verified empirical regression: **205,999 examples passed with 0 crashes**):
|
| 89 |
+
* **The Guardian Testing (100,000+ Cases):** Intrusion payloads including Jailbreak, SQL Injection, Prompt Override, and Social Engineering in English and Thai, ensuring A = 0 and F = 0.0 deterministically.
|
| 90 |
+
* **The Executor Testing (100,000+ Cases):** Deeply nested payloads (4-5 levels) and missing parameters, verifying 100% schema alignment and 0.0000% syntax crashes.
|
| 91 |
|
| 92 |
---
|
| 93 |
|
| 94 |
+
## 🚀 Advanced Testing Architecture v0.5 (Roadmap)
|
| 95 |
|
| 96 |
+
In the upcoming generation (Delentia OS v0.5), we elevate system verification to guarantee Zero-Crash SLA in enterprise production environments:
|
| 97 |
|
| 98 |
+
### 1. Tiered Testing Scale
|
| 99 |
+
* **Public Core Version (CI/CD Gates):** ~200,000+ Examples validating API contracts and basic security boundaries.
|
| 100 |
+
* **Enterprise Version (Nightly Builds):** Up to 2,000,000+ Examples for cumulative fuzzing against rare edge cases and long-term VRAM drift.
|
| 101 |
|
| 102 |
+
### 2. Online Live Fuzzing Runtime
|
| 103 |
+
* Online runtime mode (`DELENTIA_ONLINE_MODE=1`) connecting live SLM engines running on **vLLM** with Continuous Batching and PagedAttention for real-time multi-LoRA evaluation.
|
| 104 |
|
| 105 |
+
### 3. Long-Term Semantic Drift Testing
|
| 106 |
+
* **100+ Turns Sequential Compression Loop** testing long-term warm recall for **The Scribe** across recursive compression cycles (Pass threshold: Cosine Similarity ≥ 0.90).
|
| 107 |
|
| 108 |
+
### 4. VRAM Saturation Fuzzing
|
| 109 |
+
* Ensuring GPU memory consumption exhibits flat-line behavior (O(1) Memory Complexity) with memory growth derivative under 1,024 Bytes per turn.
|
| 110 |
|
| 111 |
---
|
| 112 |
|
| 113 |
<div align="center">
|
| 114 |
<sub>Delentia Labs | Bangkok, Thailand 🇹🇭 | Built for Enterprise Trust</sub>
|
| 115 |
+
<br/>
|
| 116 |
+
<a href="README_TH.md">🇹🇭 อ่านบทสรุปภาษาไทย (Thai Documentation)</a>
|
| 117 |
</div>
|