|
|
--- |
|
|
license: mit |
|
|
tags: |
|
|
- modular-intelligence |
|
|
- reasoning |
|
|
- structure |
|
|
- transformers |
|
|
- experimental |
|
|
base_model: openai-community/gpt2 |
|
|
pipeline_tag: text-generation |
|
|
language: en |
|
|
--- |
|
|
|
|
|
# Modular Intelligence |
|
|
|
|
|
Modular Intelligence is a lightweight reasoning framework built on top of a language model. |
|
|
It provides **Modules** (task-specific lenses), **Checkers** (second-pass reviewers), **Contracts** (structured output sections), and optional **Routing** (automatic module selection). |
|
|
|
|
|
The base model is GPT-2, but the architecture is model-agnostic—any LLM can be plugged in. |
|
|
|
|
|
--- |
|
|
|
|
|
## Features |
|
|
|
|
|
### Modules |
|
|
Task-specific reasoning modes. |
|
|
Examples: |
|
|
- **Analysis Note** – explanation and breakdown of concepts |
|
|
- **Document Explainer** – summaries of contracts, policies, articles |
|
|
- **Strategy Memo** – Options → Recommendation → Risks → Next Steps |
|
|
- **System Blueprint** – workflow / system design |
|
|
- **Brainstorm** – structured idea generation |
|
|
- **Message Reply** – concise responses for emails, posts, chats |
|
|
|
|
|
### Checkers |
|
|
A second pass that evaluates: |
|
|
- correctness |
|
|
- clarity |
|
|
- missing pieces |
|
|
- contradictions |
|
|
|
|
|
### Contracts |
|
|
Every module produces a fixed output template. |
|
|
This ensures reproducible structure and reduces variance. |
|
|
|
|
|
### Router |
|
|
Optional automatic module selection based on prompt classification. |
|
|
|
|
|
--- |
|
|
|
|
|
## Usage |
|
|
|
|
|
### Python |
|
|
|
|
|
```python |
|
|
from app import run_module |
|
|
|
|
|
result = run_module( |
|
|
module="StrategyMemo", |
|
|
prompt="Should we expand operations to Region X next quarter?" |
|
|
) |
|
|
|
|
|
print(result) |
