Repository navigation
Certify all nearest Clifford chart steps on the round-three word (no new κ) - #5124
Merged
Mister-Meeseeks merged 1 commit intoOct 23, 2026
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The common-axis contact search leaves independent Clifford chart changes untested. This adds a complete source-bound census of all nearest charts at every unfrozen contact and at every group constrained to preserve original zero edges on Cao Yifan’s PR4610 round-three word.
All 19,458,880 moves have nonnegative paid-minus-productive energy for every saving 0 <= C <= .005. Exact signed rank histograms, independent Python witness invoices, discrete concave dominance and outward 128-bit integer intervals establish the result. All formerly free edges, changed native warps, eight owned centers and 408 distinct arbitrary helpers are included. The selected κ remains 0.00302015. This is a finite local obstruction, not a global optimum or a new supplier.
The registered replay recompiles and reruns both complete censuses, checks all 153 tiny Lagrangians, 5,661 literal independent-axis pairs, 1,143 sequential compositions, three corruption controls and a positive descent fixture. It also replays final rank invoices for 1,960 exploratory plateau moves; those walks found no improvement and are not exhaustive.
Validation: registered replay and isolated optimized-Python replay pass; exact committed-byte/source-history preflight and official export integrity pass; a read-only Git-free isolated replay reproduces the full certificate. Hosted run 39935262575 failed in the shared onboarding fixture suite before this package replay (eight inherited failures; replay skipped). Existing PR #5123, head
95f875490c9d8d198ac594991962c4b2ebc8e017, addresses those exact fixture failures. No hosted package pass is claimed. The branch is based on refreshed canonical main20a1e31bc6009b9b3f4d084bbaadc6350e34fcba, with source inputs pinned to the original assignment base. No selected record or frozen predecessor file changes.Proof, exact scope, reproduction commands and open next questions are in
research/independent-clifford-expansion/. Joint multi-group replacements, changed scalar contacts and new native/owned protocols remain open; continued higher-κ work will change those premises.Credit Cao Yifan/PR4610, Dugongue/PR4654, the prior general Gaussian contact and phase-rank-debt packages, Cinna Davis’s PR4470/controller lineage, zxv, Jacob Alan Sussman and the retained source contributors. Prepared for Cinna Davis / 1337isnot1337 with substantial OpenAI Codex assistance; no subagents. Apache-2.0.
Canonical investigation and changed-contract continuation: #5125 .