Upload README.md with huggingface_hub

GSPC takeover densify 2026-09-07: living GET /api/gspc + Hub 885/885/0 + never certify

Wave E: additive OpenAPI/CDP apex cite (22·22·0 · x402 live · settled_usdc null)

docs: densify capability + x402 discovery (22 axis · 22 measured)

card v2.1: de-pin the npm version so the card cannot go stale, add the Python reader and the dated snapshot DOI

card v2: live-board printer, lid, DOI + citation, manifest.jsonl, viewer configs (2026-09-04)

enable the dataset viewer: declare configs for the frozen bank

card: BLUEPRINT lid beside 22 measured (14 fleets · 3 public leaders · 8 fact runs; TIE is TIE)

card: cite live GET totals.public_count (22·22·0); drop frozen MEASURED/UNMEASURED list

Reconcile README to canon 22 axis · 22 measured (14 model + 8 fact)

A5/P2.10: status-word stamp + dynamic GSPC badge + verify URL (measurement, not certification)

leftover eval.yaml: cite live GET https://councilof.ai/api/gspc. No frozen MEASURED table. Printer, not mill. Never certify. (#3)

bank card: frozen split / UNMEASURED catalog. Not a certificate. Art 50 EUR-Lex dates.

Add eval.yaml so Hub Benchmarks can run this GSPC bank

remove inspect-ai eval.yaml — GSPC forbids model-as-judge; will not fake Inspect to chase verifyToken

eval.yaml: Hub Benchmarks run OUR banks. Not certification. No VALID card → no Community Eval.

GSPC badge: 22 axis · 15 measured · 7 empty. Not certification.

Living GET printer: 22 axis · 15 measured · 7 empty (name+status only)

living GET printer: 22 slots visible, empty stay empty

sync living measurement (22-axis board)

living pointer: GET https://councilof.ai/api/gspc

docs: polish GSPC `continuity` README — 14-slot board, live ruling (#1)

docs: add intended-use + limitations (card completeness pass)

Update README.md

Update README.md

living pointer — numbers from councilof.ai/api/gspc

living board pointer, unsigned 12 Aug stamp

add measured board 2026-08-11 (RunPod 3090)

n>=30: 33 post-quantum items to give the discriminating axis an interval

n>=30: 33 post-quantum items to give the discriminating axis an interval

Add configs, pretty_name and the honesty register

Kaggle Benchmarks item set — deterministic, anti-gaming verified

initial commit