Repository navigation
Conversation
Add the scoped circuit, full scalar replay, independent integer enumeration, source pins, exact receipts and adverse controls. Signed reuse saves 926 forward real-component primitives over same-domain positive-read reuse with accepted PR696 copy resets; no backward increment survives the stronger comparator. Prepared for AmazedBot with substantial OpenAI Codex assistance. No independent human review or full multiplication-theorem verification is claimed.
Record exact low-end precision for the original and signed lowerings (forward 3 -> 4 fractional bits, backward unchanged) and cross-check work depth in the second integer engine. Every control must now fail with its stated reason. State the live-work interface and precision caveats up front, describe the two engines as same-author implementations, and trim process notes from CLAIM/VALIDATION. Circuit, routes and counts unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RGiT1dsdiBHQMR8qhHe2RZ
Add admission.py and revise_invoice.py. On a passed frontier626 replay they check that the CK native inputs are byte-identical, that coefficient work is dedicated physical stock, re-issue the retained invoice with the signed CK counts (every strict check passes), confirm the CK fractional depth (8) and global prefix (286 bits) are unchanged, and rerun the retained assembly and admission controls, reproducing kappa exactly. Only the CK contract and local recount SPEC pins are replaced. Receipt pinned as ADMISSION.json; docs updated. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RGiT1dsdiBHQMR8qhHe2RZ
Contributor
|
100-PR batch review at This remains open after scoped triage. Next review step: Local signed coefficient lowering includes a whole #626 admission claim; authenticate the reused full-replay payload and altered invoice before whole-head incorporation. This is a specific continuation point, not a rejection or a request to redo already recorded checks. Completed local evidence is retained separately from outstanding whole-head obligations. Batch review, attribution and evidence. Thank you for the contribution. |
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
3 participants
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.
Finite complex-supplier construction on the historical #626 CK provider; no new kappa claim.
Saves 926 forward real-component primitives beyond same-domain positive-read reuse with accepted PR696 copy resets. No additional backward saving.
Admitted into a fresh PR626 replay; kappa unchanged. After
scripts/verify_frontier626.pypasses,admission.py:51110230636731/31250000000000000exactly.Only the CK contract and local-recount SPEC pins are replaced; kappa, profile, hash and control-count pins are kept. Selected scalar recount: 18,263,888 → 17,530,324 real-component primitives (includes predecessor savings; 926 are due to signed reads). Not a contribution to the current #799 X4 frontier.
Base: f567884. Reproduce:
Also verified: 725 columns, 1,080,288 route cases, 15 negative controls (each must fail with its stated reason), exact fractional depth cross-checked by a second integer-only implementation, six package tests, and selected-record integrity. Both implementations are AI-written and share the cost model and inputs. All-size, realization, hosting, precision/tape and analytic obligations remain inherited.
Full evidence, source pins, predecessor credit and limitations: research/ck-signed-coefficient-routes/README.md
Prepared for AmazedBot with substantial OpenAI Codex assistance; precision check, control tightening and PR626 admission added with Claude. No independent human review is claimed.