299B Helix × GPT-2 ▶ RECORDED RESULTS — real gate outputs, not live

Helix — the verifiable execution layer. One page.

The claim: real, unchanged GPT-2 (124M → 1.5B) and a 2024 Llama-architecture model run on a software stack rebuildable from 299 hand-typed bytes, with outputs matching an independent oracle token-for-token (25/25) — gated fail-closed, reproducible, attested.

THE CHAIN

299 bytes → a compiler → 8 GPU kernels → verified AI

hex0299 B · hand-typed seedsha 9837db12… kovcself-hosts · K2==K3==K4 · 0992dddd… 8 kernelsPTX 44,019 B GPT-225/25 vs oracle
RECORDED RESULTS

Every model, gated

modelargmaxmax score difftokens
GPT-2 124M · 12 Lid 262 exact2.59e-0425/25
GPT-2-Large 774M · 36 Lid 262 exact3.8e-0525/25
GPT-2-XL 1.5B · 48 Lid 262 exact4.4e-0525/25
SmolLM2-135M · 30 L · Llama archid 260 exact4.9e-05 / 49,15225/25
WHY IT MATTERS

Audit instead of trust

FOR AI BUILDERS powered by Helix

Bring your weights: the execution layer beneath your model becomes fully traceable — same 8 kernels from 124M to 1.5B, zero new ops at scale.

FOR AUDITORS

One command (scripts/reproduce_trust.sh, ~1 min, CPU-only) rebuilds the chain from raw and asserts the anchors: 9837db12 · 0992dddd · 84363adb.

HONEST EDGES

fp32 · to PTX not SASS · single sm_86 GPU · ≈10 s/token live by design · base models, not assistants · oracle shares the spec.

Contact: ajdemarco10@gmail.com · linkedin.com/in/anthony-demarco10 · github.com/Questeria/helix  —  Web: 299bytes.com

Honest residuals: fp32 · verified to PTX, not SASS · single GPU (sm_86) · base models, not assistants · the oracle shares the model's spec. Every number on this page is a recorded, committed result. start here · guided run · expert · proof · models