# AH170 Governance Framework
This repository contains a bilingual (Malay + English) reasoning & governance dataset built from real-world business, finance, and decision-making scenarios.
The focus of this framework:
- decision guardrails (avoid risky / irrational choices)
- risk vs trade-off thinking
- stop-loss & fallback rules
- constraints awareness (cash, trust, legal, capacity)
- structured sequencing (0–72h → 7–14d → 30d actions)
Modern LLMs are good at answers, but often weak at:
- handling conflicting constraints
- defining clear decision rules
- knowing when to stop, or what to do if Plan A fails
This dataset trains models to think with governance discipline instead of “nice-sounding advice”.
- 28JSON overview & value notes
- LLM benchmark notes (Gemini / GPT comparison)
- governance scoring rubric
- 14-scenario public sample
Full dataset is private. This repo only shows structure & preview.
You are welcome to read, discuss, and evaluate the structure.
For partnerships, research collaboration, or licensing enquiries:
Email: ahadi6617@gmail.com
More modules & documentation will be added gradually.