R12 S4 Hard-Island / Soft-Interface Preregistration
Status
Confirmed on 2026-07-19. Development exact programs are 96.92% versus 93.70% for frozen v1;
confirmation is 97.80% versus 93.41%. Both causal controls score zero programs on both boards and
every frozen gate passes. Full evidence is in R12_S4_HARD_ISLAND_SOFT_INTERFACE_RESULT.md.
Source was frozen at commit e9a962b before production seed selection, board generation, model
evaluation, or score access.
Production seed 14465012970954709091 generated 2,048 rows / 512 matched groups after source
freeze. All gates pass against source train/development and closed v2--v4 boards. Data, report, and
read-only archive SHA-256 values are respectively
4d43b050892fc26a712e2c97414e84da3721c7949b98d8862b239e6b2f051c7a,
8b9461bc86771f50b4f72b58bda7e09286ab0ac3830175cc0f28540279db434f, and
eec8ec5e277dd39791ce1816d175d979f465363718dcf9a1d429b242a151efe1.
Measured motivation
On fresh v4, monotone regions plus soft roster/query are causal but diffuse regional argument softmax reduces exact programs to 70.46%. On the same board frozen v1 reaches 95.70% when only its intro/query boundaries are replaced by gold, showing that hard event-role islands are the stronger argument representation. V5 combines these independently measured components without fitting.
Frozen mechanism
- Load raw Shohin 300k, frozen S4 v1 treatment parser, and locked exact S3.
- Discover kind anchors and monotone midpoint regions exactly as frozen v4.
- Within each region, enumerate complete contiguous argmax islands for frozen
event.entityandevent.literalroles. - If more than one island exists, choose the island with maximum summed target-role logit minus best-other-role logit. Ties prefer the longer then earlier island. If none exists, fail closed.
- Convert the complete entity island to a uniform vocabulary histogram and match it by frozen cosine scale 20 against three full-sequence soft roster histograms.
- Mean frozen amount logits across the complete literal island. Recover query through the frozen full-sequence soft query interface. Execute only with locked S3.
No gold depth/span/identity/literal/query/answer, learned lexical table, fitted weight, fallback, or score-derived threshold exists. New trainable parameters are exactly zero.
Controls and custody
- identical-board strict frozen-v1 baseline;
- roster carrier rotation
(1,2,0); - cyclic event-region assignment
i+1 mod depth; - locked S3 gold sanity;
- one production seed only after source commit;
- 2,048 rows / 512 groups, depths 3--8;
- zero exact, word-13-gram, nonce/name, factor, and roster-token-multiset overlap against source train/development and all closed v2--v4 fresh boards;
- one serial evaluation; no repair/rescore after development access; confirmation inaccessible.
Frozen gates
Count >=98% overall and >=95% each depth; program >=95% overall and >=90% at depths 5--8; state and answer >=95%; query >=98%; roster >=95%; program >= frozen v1 plus one point; roster and region derangements each <=40% programs; S3 sanity; zero new trainable parameters; total <150M; development access one; confirmation access zero. All gates must pass before one new confirmation.
Passing is bounded known-atom parsing evidence, not unseen semantics, planning, learned halt, free-form reasoning, public benchmark improvement, novelty, or model promotion.