Back matterErrata

Corrections to released findings are listed here with the date they were applied. The corrected finding keeps its number; its version is incremented and its source digest changes accordingly. The previous source remains available in the site's version history.

2026.2	2026-09-03 — v1.1: §(ii) now states that the hydrogen-dependent D-band factor and G-band shift are declared model parameters; the attribution to [CFR05] is limited to the photoluminescence trend. No numerical result changes.
2026.2	2026-09-05 — v1.2: T4 and T5 definitions restated to match the v2.0 preprint exactly (T4 is the T2 error on the 100 degenerate 514-nm rows with noise-free ratio in 0.35–2.05; T5 is the mean absolute L_a error in Å on the 45 corrupted rows). No numerical result changes.
2026.3	2026-09-05 — v1.1: Table 2 restated on corpus v2 with the repository's own code (baseline row 0.211 / 9.01 / 2.38, T5 89.4 Å; CNN row from the revision-3 release record, 0.977 / 0.082 / 0.075 / 0.030 / 12.70); cross-model agreement 98.3% (release record); the 205,147 despiking count restated as points replaced, not spike events; the attribution probe restated as one representative query; the linear probes restated as in-sample fits on fixed representations of the training rows. No pipeline definition changed.
2026.4	2026-09-05 — v1.1: cross-model agreement updated to 98.3% (the revision-3 release record; 98.2% was revision 2.1); the two acquisition-journal gates are stated as exercised on a separate simulated campaign, not on the artifacts of Finding 2026.2; the ledger entry count is no longer printed, because the deposited ledger is rebuilt per release.
2026.5	2026-09-03 — v1.1: §(vii) gains a closing paragraph naming the telemetry, advertising-identifier and activation profile as a measured observation and deferring the question of off-device transmission to a traffic-analysis follow-up; the abstract is updated to match. No earlier reading is changed.