Skip to content

Hinedes/ARW

Repository files navigation

ARW — Asymmetric Read/Write

One-way gradient gate for zero-forgetting continual learning via orthogonal subspace projection.

When a model is fine-tuned on a new domain, ARW lets it read all prior knowledge (forward pass sums all domain projectors), but only writes gradients to the new domain's own orthogonal shell (backward pass locked to one projector). The result: the base model remains immutable, while the new domain learns rapidly.

Status

  • Mathematical formulation (split projectors)
  • GPT-2 proof-of-concept: Domain 0 retention +0.000, Domain 1 PPL 8.643
  • Overnight cluster experiment
  • Jailbreak-resistance test
  • Paper draft

Related Projects

  • Proteus — Matryoshka Subspace Freezing (first implementation of the core-freeze idea).
  • Potential Well Project — Orthogonal subspace partitioning (precursor; identified the Orthogonality Paradox that ARW resolves).

Quickstart (coming soon)

git clone https://github.com/Hinedes/arw.git
cd arw
# Install dependencies
pip install torch transformers datasets
# Run ARW experiment
python train_arw.py --help

About

Asymmetric Read/Write (ARW) — one-way gradient gate for zero-forgetting continual learning using orthogonal subspace projection.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages