Datasets:

PureOne
/

SORILUN-RSI

# Claim and evidence ledger

| Claim | Evidence | Boundary |
|---|---|---|
| High-order inverse transport is exact | Complete proof; `tests/resonant_verify.py` | Formal power series, stated precision/order hypotheses |
| Hasse supports can enlarge the exact update window | Complete proof; 376 checks | Sufficient certificate, not a necessary condition |
| Boundary parent depth is sharp | Complete proof; 181 checks, including nonzero witnesses | Ordinary coefficient-prefix model; not arbitrary encodings |
| Cartier transport commutes | Elementary proof and exact tests | Established underlying mathematics; no claim of inventing Cartier naturality |
| Typed reconstruction is lossless | Formula, proof, exact tests | Only the specified resonant source class; parent carrier remains necessary |
| The learned implementation can cheapen its own fixed-candidate validation | Same-child interventions and cost records | Supplied candidate grammar; no discovery of better scientific ideas |
| Full inverse state can be updated continuously | 24 successive verified updates | Finite executed stream, explicit high-order changes |
| Persistent compiler commits improve the restricted service | 3 commits across 18 requests | Static normal form has the same endpoint |
| Solver work and measured time are repaid in the resonant service | `results/resonant_service.json` | Frozen uncompiled comparison, declared timing boundary |
| Broad arithmetic timing superiority | Not established | Strong NTT and static controls retained |
| General autonomous RSI | Not established | Restricted learning/operator language only |
| Historically unprecedented theorem or architecture | Not established | Targeted literature audit is not a priority proof |
| Independent replication or peer review | Not completed | Internal execution and hand-written proofs only |
| Pre-execution investment reservation | 305 tests of the new helper | Old experiment settles retrospectively; no retroactive relabeling |
| Timing evidence controls false promotions | Conditional supermartingale proof | Win-probability null, conditional assumptions; not mean-time or drift guarantee |

Potential if confirmed in independent review: a specialized mathematical contribution with a reproducible resource-recursive application. A field-defining general-RSI claim is not supported by the demonstrated scope, regardless of review status.

	# Claim and evidence ledger

	\| Claim \| Evidence \| Boundary \|
	\|---\|---\|---\|
	\| High-order inverse transport is exact \| Complete proof; `tests/resonant_verify.py` \| Formal power series, stated precision/order hypotheses \|
	\| Hasse supports can enlarge the exact update window \| Complete proof; 376 checks \| Sufficient certificate, not a necessary condition \|
	\| Boundary parent depth is sharp \| Complete proof; 181 checks, including nonzero witnesses \| Ordinary coefficient-prefix model; not arbitrary encodings \|
	\| Cartier transport commutes \| Elementary proof and exact tests \| Established underlying mathematics; no claim of inventing Cartier naturality \|
	\| Typed reconstruction is lossless \| Formula, proof, exact tests \| Only the specified resonant source class; parent carrier remains necessary \|
	\| The learned implementation can cheapen its own fixed-candidate validation \| Same-child interventions and cost records \| Supplied candidate grammar; no discovery of better scientific ideas \|
	\| Full inverse state can be updated continuously \| 24 successive verified updates \| Finite executed stream, explicit high-order changes \|
	\| Persistent compiler commits improve the restricted service \| 3 commits across 18 requests \| Static normal form has the same endpoint \|
	\| Solver work and measured time are repaid in the resonant service \| `results/resonant_service.json` \| Frozen uncompiled comparison, declared timing boundary \|
	\| Broad arithmetic timing superiority \| Not established \| Strong NTT and static controls retained \|
	\| General autonomous RSI \| Not established \| Restricted learning/operator language only \|
	\| Historically unprecedented theorem or architecture \| Not established \| Targeted literature audit is not a priority proof \|
	\| Independent replication or peer review \| Not completed \| Internal execution and hand-written proofs only \|
	\| Pre-execution investment reservation \| 305 tests of the new helper \| Old experiment settles retrospectively; no retroactive relabeling \|
	\| Timing evidence controls false promotions \| Conditional supermartingale proof \| Win-probability null, conditional assumptions; not mean-time or drift guarantee \|

	Potential if confirmed in independent review: a specialized mathematical contribution with a reproducible resource-recursive application. A field-defining general-RSI claim is not supported by the demonstrated scope, regardless of review status.