Files

T

leetcrypt 47b2937789 SS2: raw-data freeze — confirmatory battery COMPLETE 180/180, checksummed

Battery output/sor-confirmatory/20260720T060132Z/confirmatory-data finished:
180 cell dirs, each metrics.json + manifest.json + non-empty circuits/ with
per-hop pcaps. Integrity PASS: 0 missing pcaps (27000 pcap + 9000 events.jsonl);
seed provenance verified 180/180 (seed == SHA256(S0=20260719 || cell_id ||
run_index)); R=30 held exactly across all 6 frozen design cells (RQ1{off,on,
on+padding}, RQ2{1house-N,bridge-federated,directory-federated}). No aggregate
battery-results.json (expected, NOT fabricated — 180-cell count is the frozen
completion proxy).

Immutability anchor: SHA256SUMS.txt over 36361 raw artifacts (sorted rel paths);
provenance.json (S0 + seed formula, executor SHA ce68d41, engine=docker
image_digest=null@runtime, timestamps, completion proxy); INTEGRITY-REPORT.md.
Anchors force-added (/output/ gitignored → raw data stays out of git). Raw
metrics.json IMMUTABLE — not recomputed or edited.

SS2 touched structure/checksum/provenance ONLY — no RQ1/RQ2 inferential
statistic (AUC/H/CI) computed or inspected. Next: SS3 frozen-§6 pass.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-07-20 22:35:15 -07:00

2.5 KiB

Raw Blame History

SS2 raw-data freeze — integrity report

Battery: output/sor-confirmatory/20260720T060132Z/confirmatory-data Frozen prereg SHA-256: f22331a72e0d0ccf38b787e63acabbe9d666456ec76076787a6d545c3193425b Frozen §4 sample: R = 30 runs × C = 50 circuits per cell. Completion criterion (frozen): 180 cells each carrying metrics.json — MET.

Integrity assertions (all PASS)

Check	Result
Cell dirs	180
Each has `metrics.json`	✅ 180/180
Each has `manifest.json`	✅ 180/180
Each has non-empty `circuits/`	✅ 180/180
Missing per-hop pcaps (hop0/1/2 × 9000 circuits)	0
Seed provenance `seed == SHA256(S0‖cell_id‖run_index)` (S0 = 20260719)	✅ 180/180 match, 0 mismatch

Design-lattice mapping (R = 30 balance held)

RQ3 is not in the frozen lattice (dropped for the G4+RQ1+RQ2 lead paper). The 6 design cells (3 RQ1 arms + 3 RQ2 arms) each carry exactly R = 30 runs:

cell_id	runs
RQ1/topo=1house/selector=static/bridge=off	30
RQ1/topo=1house/selector=static/bridge=on	30
RQ1/topo=1house/selector=static/bridge=on+padding	30
RQ2/bridge=off/selector=static/topo=1house-N	30
RQ2/bridge=off/selector=static/topo=bridge-federated	30
RQ2/bridge=off/selector=static/topo=directory-federated	30

RQ counts: RQ1 = 90, RQ2 = 90. Selector = static; churn_schedule_id = none (RQ1/RQ2 use no churn).

Immutability anchor

SHA256SUMS.txt — SHA-256 over 36361 raw artifacts (27000 per-hop pcaps + 9000 per-circuit events.jsonl + 361 json = 180 metrics + 180 manifest + 1), sorted by relative path. Manifest-of-manifests digest: 3fb67c7a9253c7f61f25a8432224decd85cff1b4baee8f6a1b60baa936707922.

Raw metrics.json are immutable — not recomputed or edited in this stone.

Provenance notes

No aggregate battery-results.json was written (expected; not fabricated). Cell count 180/180 is the completion proxy.
engine.image_digest was recorded null at run time (sor-hop image not digest-pinned by the executor); containment engine != local was asserted per-hop by ForwarderPlan at collection time.
Battery window (mtime proxy): start 2026-07-20T06:11:41Z → end 2026-07-21T05:24:09Z (≈ 23h).

Blinding

Blinding is lifted for code-on-real-data. SS2 touched only structure/checksums/provenance — no RQ1/RQ2 inferential statistic (AUC / H / CI) was computed, aggregated, or inspected. The first real inferential numbers come from the frozen §6 pipeline in SS3, in one auditable pass.

2.5 KiB Raw Blame History Unescape Escape