Neural network training stability

Detect. Attribute.
Intervene.

Bendex monitors your training run in real time, identifies which layer is failing, and corrects it automatically — before your run is lost.

100%Detection rate
0%False positive rate
90%Task recovery rate
7Validated architectures
30Seed benchmark

How it works

Three steps.
Zero crashes.

Bendex runs alongside your training loop with a single observe() call per step. No wrappers, no rewrites.

01 — Detect

Weight divergence trajectory curvature

Bendex tracks the geometric trajectory of your model's weights relative to a frozen reference. The discrete curvature κ_t spikes at instability onset — well before your loss curve shows anything.

02 — Attribute

Deepest-early-cluster attribution

When a trigger fires, Bendex identifies which specific module deviated first using per-layer z-scored divergence histories. You know exactly where the problem originated.

03 — Intervene

Three-phase automated correction

Bendex freezes the suspect module, reduces learning rate, and monitors recovery. If the run doesn't stabilize, it escalates. Your training continues — automatically.

Benchmarks

Compared against
every alternative.

Bendex was benchmarked against loss spike detection, gradient norm monitoring, and patience-based early stopping — the approaches most teams currently use.

SystemDetectionFalse positive rateRecoveryAttributes moduleIntervenes
No monitor0%0%0%NoNo
Loss spike100%80%0%NoNo
Gradient norm90%50%0%NoNo
Patience / early stop100%50%0%NoNo
Bendex100%0%90%YesYes

Research

Five papers.
One framework.

Bendex is the applied proof of a theoretical program in information geometry. The same Fisher manifold underlying Bendex's zero false positive rate also derives the fine structure constant α to 8 significant figures — a blind prediction from pure geometry.

Paper 01
Informational Curvature: A Geometric Framework for Structured Information
Figshare · 2026 · CC BY 4.0
Paper 02
Informational Stability: A Geometric Framework for Structured Information Over Time
Figshare · 2026 · CC BY 4.0
Paper 03
Reflexivity: A Geometric Framework for Self-Referential Dynamics
Figshare · 2026 · CC BY 4.0
Paper 04
Experience: Internal Availability of Reflexive Dynamics
Figshare · 2026 · CC BY 4.0
Paper 05
Specialization: Stabilized Reflexive Loops and the Constants of Nature
Figshare · 2026 · α to 8 sig. figures

Start protecting
your training runs.

Bendex is in early access. Fill out the form and we'll be in touch within 48 hours with onboarding instructions and licensing details.

Questions? Email 9hannahnine@gmail.com

We respond within 48 hours. No spam.