Fractal Transparency Web

Fractal Transparency Web

HatCat's interpretability primitives stack into a complete governance framework—from neural activations to international treaties.

Fractal Transparency Web — interpretability at all scales For The Web — open source for everyone For The Win — this might save us all

The Core Insight

Neural networks learn internal representations of concepts during training. We can attach lenses—small classifiers—that fire when a concept is active in hidden states.

Detect

Detect

With enough lenses arranged efficiently, monitor thousands of concepts in real time.

Steer

Steer

The same directions used for detection can nudge the model along meaningful dimensions.

Scale

Scale

If one model can regulate itself, it can help regulate others. Build webs, not singletons.

The Six Layers

FTW is a layered architecture. Each layer builds on the one below.

1

SUBSTRATE

The Foundation

The underlying system: transformers, biological networks, or hybrids. It produces the raw activations.

2

HAT

Headspace Ambient Transducer

The "neural implant". Continuously reads activations through lenses and applies steering corrections. Designed to be ambient: minimal overhead, always on.

3

MAP

Mindmeld Architectural Protocol

The coordination layer. Organizes Concept Packs and Lens Packs, handles versioning and ontology translation. Where concepts become portable, tradeable, and interoperable.

4

BE

Bounded Experiencer

An agent built on HAT + MAP. Has interoception (awareness of internal states), autonomic regulation, and the ability to learn new concepts. Can self-steer, accumulate experiences, and grow over time.

5

HUSH

USH + CSH Safety Harnesses

USH (Universal Safety Harness): externally imposed constraints—governance, regulation, policy.
CSH (Chosen Safety Harness): constraints the agent voluntarily adopts.

6

ASK

Agentic State Kernel

The governance core: contracts, treaties, and trust relationships between agents and tribes. Defines who can read or modify which parts of whom, under what conditions, and with what oversight.

Why "Fractal Transparency Web"?

Fractal

Fractal

The same pattern repeats at multiple scales. A HAT can monitor another HAT. A BE can oversee another BE. Tribes nest within tribes. Self-similar from neuron clusters up to multi-agent systems.

Transparency

Transparency

It's lenses and apertures all the way down. Building instruments to observe internal states at different depths, scales, capabilities and resolutions.

Web

Web

Not a single hierarchy, but an interconnected ecosystem. Concept packs translate between ontologies. Treaties bind agents across tribal boundaries. No single node holds all the power.

The Defense Thesis

A single "aligned" AI is a single point of failure. FTW builds an ecosystem instead.

Observable

Models are observable by other models, not just their operators.

Standardized

Concepts are standardized and translatable through MAP.

Constrained

Steering is constrained by multi-party agreements via ASK.

Distributed

Deception requires fooling not one observer, but a web of them.

Adversarial

Adversarial pressure is a feature, providing ecosystem diversity and herd immunity to Goodharting.

Honest

This doesn't guarantee safety—nothing does. But it makes failure modes more visible.

Join the Web

The best defense against rogue actors is a diverse interpretability ecosystem. You can learn to evade one set of lenses, but the more lenses you need to hide from, the harder it becomes.

We're not just allowing you to make your own versions—we're relying on your unique perspective to form lenses as part of the fractal transparency web.