Quant & research

Auditable attention paths. Measured energy. Public packs.

Infrastructure for deterministic, receipt-attested, energy-aware AI compute — with numbers you can open on GitHub.

Core capabilities

TRADE residual stacks

Device-resident GPU stacks for multi-layer residual compute. Public 7B-class path: 28 layers × h=3584 on H100 NVL with multi-run thr and sustain-only J/token.

AUDIT & receipts

Where the product lane demands it: cryptographic receipts, free-ride residual checks, and null-space behavior under load — not just marketing “determinism.”

WNSM free-ride

Null-space payload bus under real CUDA load. Public pack: free-ride vs side-channel H2D, null residual ~1e-8 class, single-layer null-inject drift 0 in the stack test path.

Long-context O(N) memory

Waller streaming state scales ~O(N) in memory vs dense scores ~O(N²). Public ladder through 32k (and analytical 131k memory reduction).

Convert + serve path

HF → Luxi native convert for 7B-class weights; serve and TRADE examples for operator evaluation. Implementation lives in the engineering repo; proof is public on LuxiDemo.

Deterministic math demo

Standalone binary: JSON expressions in, results + SHA256 out. Useful for integration smoke tests — secondary to the inference-energy thesis on this site.

Where we win vs where we are honest

AxisStatusPublic proof
7B-class board J/token @ seq≥128 ~0.63 J/tok · ~403 tok/s h100-7b-class-TRADE
Short-seq thr vs Flash Flash-class may win (published H2H loss) h100-stack12-H2H
Free-ride / AUDIT under load Differentiated thesis h100-WNSM-free-ride
Long-ctx memory scaling O(N) vs O(N²) h100-LONGCTX-scaling
Not HF chat SOTA. The 7B-class TRADE path maps architecture into TRADE kernels (e.g. GQA/SwiGLU into the GELU-MHA TRADE path). Use it for energy / stack evaluation, not as a drop-in chat quality claim.