Perturb
Harden Your AI Against Invisible Attacks
Perturb probes production AI models 24/7 with adversarial attacks — delivers vulnerability heatmaps.
The simplest explanation
Like a red team that never sleeps: constantly probing AI systems to find weaknesses.
Everything you need to understand Perturb
No jargon. No whitepapers. Just the facts.
The Problem
AI models are silently vulnerable to adversarial inputs; single-pixel or noise perturbations can flip state-of-the-art model outputs, and centralized red-teaming is narrow and infrequent.
What Perturb Does
Perturb probes production AI models 24/7 with adversarial attacks — delivers vulnerability heatmaps.
The Edge
Global network of miners attacking models 24/7 with rewards aligned to finding real exploitable vulnerabilities rather than synthetic ones; on-chain proof of evaluation for compliance.
What's the closest mainstream equivalent?
The best way to understand a new technology is to compare it to something familiar.
Global network of miners attacking models 24/7 with rewards aligned to finding real exploitable vulnerabilities rather than synthetic ones; on-chain proof of evaluation for compliance.
How does Bittensor make this possible?
Perturbuses Bittensor's incentive layer to build something no single company could run alone.
Miners compete
Participants (called miners) on the Perturb subnet compete to produce the best outputs for ai security / adversarial robustness tasks. Anyone with the right hardware can join.
Validators score the work
Validators continuously evaluate miner outputs against objective benchmarks. The best performers rise, the worst are replaced. There's no human committee — the protocol decides.
Rewards flow to the best
Miners are paid in the Perturb alpha token in proportion to how good their work is. This creates a continuous competitive pressure that drives quality up and cost down — structurally, not just as a promise.
The whole network benefits
Because every participant is aligned toward the same goal — producing the best ai security / adversarial robustness results — the network improves continuously without requiring a central team to manage it.
Frequently Asked Questions about Perturb
What is Perturb on Bittensor?
Perturb is Subnet 26 (SN26) on the Bittensor network — a decentralized AI protocol built on the TAO blockchain. Perturb probes production AI models 24/7 with adversarial attacks — delivers vulnerability heatmaps.
What problem does Perturb solve?
AI models are silently vulnerable to adversarial inputs; single-pixel or noise perturbations can flip state-of-the-art model outputs, and centralized red-teaming is narrow and infrequent.
How is Perturb different from Centralized AI red-teaming / model-security vendors (HiddenLayer, Robust Intelligence, Protect AI)?
Global network of miners attacking models 24/7 with rewards aligned to finding real exploitable vulnerabilities rather than synthetic ones; on-chain proof of evaluation for compliance.
What is the Perturb token?
The Perturb subnet has its own alpha token on Bittensor's dTAO system. It trades in the Bittensor liquidity pool and its price reflects market demand for the subnet's services.
Is Perturb a good investment?
AlphaGap tracks Perturb using its aGap score — a composite of development activity, token flow, and social signals. This page is for informational purposes only and is not financial advice. Always do your own research before making any investment decisions.
Want real-time intelligence
on Perturb?
AlphaGap tracks signals, whale flows, developer commits, and the aGap score for every Bittensor subnet — updated continuously. Find the alpha gap before everyone else.
Free preview · Pro from $29/mo · Premium from $49/mo
