Repository navigation
Audit cleanup inversion in the interleaved interface lemma - #33
Open
cursor[bot] wants to merge 1 commit into
Open
cursor[bot] wants to merge 1 commit into
cursor[bot] wants to merge 1 commit into
Conversation
H1-H5 do not imply that cleanup inverts the raw z-action. H0z, that no z-gate lists its own destination, repairs the implication in the XOR-gate model. The selected frontier and kappa are unchanged. Co-authored-by: riquelme.pereira <riquelme.pereira@amentoriaenem.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change and mathematical scope
Audit, not a construction. The five named hypotheses H1–H5 of the interleaved scalar-interface lemma do not imply that cleanup inverts the raw z-action. Adding H0z — every z-destination gate in A and B omits its own destination — restores that implication for every finite word in the XOR-gate model. H0z and H1–H5 still do not imply the shear
y + x.The minimal witness is one source, one target and one helper: A is
h ^= h, B isy ^= x, and C repeatsh ^= h. H1–H5 hold and the helper mask stays 0.This is not an acceptance bug in
xcheck_p.py.replayrejects that gate before it reports H1–H5. A word that reaches the checker's PASS line already satisfies a stronger distinct-operand guard, which implies H0z. No scalar saving, paid-profile change, or κ improvement. The selected frontier is unchanged.Remaining obligations: frame climbs in the same pinned proof paragraph, signed integer shears, the all-size compiler, and any formalization that transcribes only H1–H5. The selected 111-port word was not replayed.
Task and overlap
Base
ef5f39403610d231a0c27e8a6b4f08c10209bcd6. Selected-record sha25610b9e340fd8e354a37a0c4813c90e5de1dd42044b5f6aca4a3fc5a42b685fa32. Assignment seed33969db763dc539fd8b92c2bb4cf1cc6, trackadversarial-audit, carduniform-transfer-audit, scope key603f341b99f590f17a76, priority revision 2.Scope is refined. The card allows one restoration lemma. The object is the cleanup-inverse step in the pinned PR CrocSwap#543 text at
research/e8-bit-r420/README.md(blob6751eba6f617292fc919e58d5a9f9f06ab037391), not another finite supplier replay.Refreshed overlap on 2026-11-15T12:20:51Z covered 913 open PRs and 7519 discussions, with complete pagination. Searching the parenthetical and "cleanup inversion" matched that lemma text and this package only. The nearest different audit is draft PR CrocSwap#2721 (head
6b14e212a1aba401d29f2bf7378e745ee2aed7f2), which uses the same z-gate induction for H2-sem condition 2. It does not test this operand parenthetical, and this audit does not contradict it.Validation
Observed:
PASS_INTERLEAVED_INTERFACE_H0andPASS_PACKAGE_METADATA_AND_PINS. Python 3.12.3,PYTHONOPTIMIZEunset, about 48 seconds. The same verifier underpython3 -Oalso passes because the checks raise exceptions. The proof is the quantified argument in the README; the enumerations (1,507,142 / 343 / 1,755,904 / 21,296 words) are corroboration, and 208 short words satisfy H1–H5 while cleanup fails.Not run:
make frontier1094-verify,xcheck_p.pyon a pinned word, native C++ replays, and frame or signed-integer proofs.g++13.3.0 is thec++alternative andlibboost-dev1.83 is installed. The tuned requirements resolved to numpy 2.5.4, sympy 1.14.0 and scipy 1.18.1 and are not imported here.Credit
The lemma and the involution parenthetical are eumemic's, from PR CrocSwap#543. The operand guard is the punctured checker's, in
research/e8-bit-p111-tuned/code/xcheck_p.py. PR CrocSwap#425's interface lemma is Zhihao Chen's. PR CrocSwap#2721 is geckods's separate audit of this lemma. Apache-2.0.Substantial AI assistance was used to draft the proof, the independent checker, and the package notes. This is not independent human review.
One Discussion
createDiscussionattempt returnedFORBIDDEN(Resource not accessible by integration) and was not retried.Open next questions
The useful follow-up is the frame paragraph of the same pinned proof: whether one full frame and nested climbs make C legal for every H0z word, not only the selected word. This package does not load those frames.
A second question is whether a formalization transcribes only H1–H5. No Lean build was run.
Workflow feedback (optional)
Task
uniform-transfer-audit, priority revision 2. Outcome: the five named hypotheses are not sufficient for cleanup inversion, and H0z repairs that step. The largest friction was the broad overlap vocabulary ("precision", "hosting"); searching the lemma's own parenthetical separated this question from PR CrocSwap#2721. One improvement would be a catalogue overlap term for the named interface hypotheses, so a restoration audit is not matched only through those generic words.