wpgit: version control and regression gates for LLMs editing neural weights directly
bigspin5050 · reddit · 2026-11-05
A Reddit author released Weight-Space Software Engineering (WSSE) and wpgit (paper + code), addressing what happens when LLMs make direct gradient-free weight edits (à la MIT CSAIL's 'Sparks of Weight-Space Programming') at scale: silent regressions, merge conflicts, and breakage after base model updates.
Core design — weight edits as text diffs:
- Each edit is a human-reviewable .weightpatch.json, hash-pinned to the base checkpoint, dependency-aware, using compact operators (e.g. a 12-number affine fold rewriting 4,803 weights), and carrying quantitative assertions that must pass before merging.
- wpgit layers on Git: build (topological checkpoint compilation), test (differential integration tests in context), check (pairwise commutativity to catch semantic merge conflicts), rebase (migrate patches to updated base checkpoints).
Validated across 17 experiments (150K–8.1M params, including CatGAN, MobileNetV2, and legacy Keras checkpoints): rank-2 active-subspace locks collapse attribute variance from 70% to 11% without retraining, and channel reclamation + exact ReLU gating performs zero-parameter circuit surgery (e.g. turning only cat irises blue).
More from Research
- USC's WACL teaches robots whole-arm manipulation while the hand stays busy — ebiyik_ · 2026-11-05
- First-ever asymptotic for E.P. Stein numbers proved with an AI collaborator, published on arXiv — arampell · 2026-11-05
- Weight-space programming's soft claims rest on 3-4 correct answers on toy tasks — maier_ak · 2026-11-05
- Weight-Space Programming: what's measured vs anecdote in MIT CSAIL's GPT-6 claims — maier_ak · 2026-11-05
- ExMR open-sources single-video full-body 3D motion capture with robot export — camenduru · 2026-11-05
- VA-TDM: Velocity Adversarial Distillation Sets SOTA for Few-Step Diffusion Models — Yihong Luo · 2026-11-05