Skip to content
485 changes: 485 additions & 0 deletions research/r437-composed-bit-independent-audit/README.md

Large diffs are not rendered by default.

1,456 changes: 1,456 additions & 0 deletions research/r437-composed-bit-independent-audit/RESULT.json

Large diffs are not rendered by default.

132 changes: 132 additions & 0 deletions research/r437-composed-bit-independent-audit/SOURCES.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,132 @@
{
"schema_version": 1,
"base_commit": "20ecf2891d80b5acf18d980b2164fb44ea7faa90",
"files": [
{
"path": "certificates/selected-result.json",
"usage": "provenance",
"git_blob": "b1fea0c94436b236cc2b7d41a7be032811b94c81",
"sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"path": "docs/research/agent-task-catalog.json",
"usage": "provenance",
"git_blob": "67f085acd86acd7d4e03922d9cf4b26bca66fed5",
"sha256": "47b42433f5a85809a3aadd29c2c339da2c6ed728e4f7762df6b88bccb7874cf0"
},
{
"path": "research/e8-bit-p111-tuned/tune/bitE8mut-p111-tuned.json.gz",
"usage": "input",
"git_blob": "85a4499bf9e4911a705d50567a2c2ef6cf23739a",
"sha256": "e3e301e62bc2b41dbf1ceac0acf3d35824ebeab1848c5a2142d4577b109da52f"
},
{
"path": "research/bit-boundary-persistence/FRAME-PATHS.json",
"usage": "input",
"git_blob": "b1862b43e1fa1ee0fb02e29b303801265638dae4",
"sha256": "0d347b9e4d8a8336a707565deceb1d505cd138c97a56831643583b83d91c3b4f"
},
{
"path": "research/bit-boundary-persistence/PROOF.md",
"usage": "provenance",
"git_blob": "560c21b5d47c2779d2b2e14b4decb1aaeba4c64f",
"sha256": "fee9059a1fb1db9ec16ce339f0dd36c29f2e6ec7103aa76c0b2296f8529bac6c"
},
{
"path": "research/e8-bit-p111-tuned/code/xcheck_p.py",
"usage": "provenance",
"git_blob": "184f5dcf94a486cc7db91ca4db5865916bf26009",
"sha256": "af3decc877e27c00001cf56f50c6c09566e8e101297ab34c7860aed96a06d80e"
},
{
"path": "research/e8-bit-p111-tuned/code/pwordl.py",
"usage": "provenance",
"git_blob": "a90bed93e227c3916aca67817794f35e24123ceb",
"sha256": "6d8fccd44508e4705b6446e0a64482a5221a3efc5663c5ca7ae90d44581ac50a"
},
{
"path": "research/e8-bit-p111-tuned/data/selection9-p111.json",
"usage": "input",
"git_blob": "e69ea66a3add357f570fc5552ab6368b3c8f9bb2",
"sha256": "4ab1ed4529683cd184987f8bc3031fb7408b9cd22c25ab9cf541a8aaf896eaf1"
},
{
"path": "research/bit-boundary-persistence/CERTIFICATE.json",
"usage": "input",
"git_blob": "56400920050c523e6e399f2455137b36c35cf528",
"sha256": "32c6cd6103006d627909cabdaf5083238c661a6582016a8b21075bd01c892eff"
},
{
"path": "research/e8-bit-carrier-skeleton/code/hub/idle-axis-hosting/hosting_core.py",
"usage": "provenance",
"git_blob": "68b536daa9e6546c0aa046934bf840aaa3c3fde3",
"sha256": "d8b90cd5d152df3c7c351230863d9e61816ad5e1f21d69bea1a86a5daad559a3"
},
{
"path": "research/e8-bit-carrier-skeleton/code/hub/e8-r740-replica32/code/engines/moment.py",
"usage": "provenance",
"git_blob": "7d697cb45203fe72d97037403d6da41f71f3b7fd",
"sha256": "81cbc864e106418fc7854322746f72180cb8110ff39f74450a3b8d55a82c1482"
},
{
"path": "research/e8-bit-p111-tuned/vendor/pr510/e8-r480-nine-stage/vendor/e8-r715-eleven-stage-whole/code/global_k.py",
"usage": "provenance",
"git_blob": "3df12bd632f1fdba825ced57110a1fee82eb2857",
"sha256": "6368196a393ac25bb2918880e4c4bd31ed733b762c27b6873ab391cbc0c5c3c8"
},
{
"path": "research/e8-bit-p111-tuned/vendor/pr510/e8-r480-nine-stage/code/assemble9.py",
"usage": "provenance",
"git_blob": "238226c568742dd5723b03eabae564a2eaa2bda9",
"sha256": "4bf6768224a3fac81ccda26df24f3763f9a4fbc35ad62bef2888595cdae519fe"
},
{
"path": "research/e8-bit-carrier-skeleton/code/bit/checkers/columns_h.cpp",
"usage": "provenance",
"git_blob": "00c5a3a5d81d083582f60f33dd2e7bea5049c37f",
"sha256": "d1297f9052be4f2496e1ff2a77e4db2c0cdbd7d460d489b4c03e50bf317238ce"
},
{
"path": "research/e8-bit-carrier-skeleton/code/bit/checkers/columns_k.cpp",
"usage": "provenance",
"git_blob": "42bb232fbec4de703e70cc13dc2f39a3b43a3a01",
"sha256": "1c344d18fcd494eeca93e259f10cafae7705780e6592f6dcb35b34752dab9bc2"
}
],
"snapshot": [
{
"local": "inputs/R437-COMPOSED-WORD.json.gz",
"source_path": "research/bit-p111-r437-gl464-470-20261018/data/r0042/R437-COMPOSED-WORD.json.gz",
"source_commit": "36ff441479ff9c7e70d83f0becccd44da50effb8",
"source_pr": 3647,
"git_blob": "c558a221dd44cbdaa732a401f9876de4c096d4a7",
"sha256": "e432182d5b62c8bbefeb871fce30317234d707e5e18d4295e65ac21f7fd8365e"
},
{
"local": "inputs/FINAL-R438-WORD.json.gz",
"source_path": "research/bit-p111-r437-gl464-470-20261018/data/r0028/FINAL-R438-WORD.json.gz",
"source_commit": "36ff441479ff9c7e70d83f0becccd44da50effb8",
"source_pr": 3647,
"git_blob": "db621440ab3f87ece927137c21f4f7fb3ba221ef",
"sha256": "5aa4d91fb0062265b25fb8f322a57b76d89caaf055b4110cf6fdb761f0b27463"
},
{
"local": "inputs/R437-RECORDS.bin.gz",
"source_path": "research/bit-p111-r437-gl464-470-20261018/data/r0042/R437-RECORDS.bin",
"source_commit": "36ff441479ff9c7e70d83f0becccd44da50effb8",
"source_pr": 3647,
"git_blob": "ac75dd07819be402dc28a84b09561bfe42f1cc80",
"sha256": "6e556dae1a2f56f3893afa9b316f6029712c6b79a3536a429f8d83cac79163e9",
"uncompressed_sha256": "caeb2eb0e417992788c32c1f5812ada60b46d1dfdba9f4ac376299b2dede13cb",
"note": "gzip (level 9, mtime 0) of the PR file; the verifier checks the decompressed SHA256"
},
{
"local": "inputs/R437-GLOBAL-TEMPLATE.json.gz",
"source_path": "research/bit-p111-r437-gl464-470-20261018/generated/GLOBAL-TEMPLATE.json.gz",
"source_commit": "36ff441479ff9c7e70d83f0becccd44da50effb8",
"source_pr": 3647,
"git_blob": "df9238596dbe8cd355a65aa170caa54f543c39cd",
"sha256": "00dee04c5372ac300b678c4ac1e6770c178923099d2d4bf01be9e4bccd2a1e72"
}
]
}
84 changes: 84 additions & 0 deletions research/r437-composed-bit-independent-audit/TASK.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,84 @@
{
"schema_version": 1,
"task_id": "independent-moment",
"priority_revision": 6,
"base_commit": "20ecf2891d80b5acf18d980b2164fb44ea7faa90",
"original_card": {
"id": "independent-moment",
"question": "Does one changed supplier or assembly claim lack a source-bound independent check of a specific new implication?",
"starting_points": [
"certificates/selected-result.json",
"research/rank7-joint-endframes/README.md",
"docs/research/research-family-status.md"
],
"methods": [
"Independently reconstruct a coordinated frame/circuit change and its single-change control, then derive the complete rank and physical ledger.",
"Implement certified moment inequalities through a different exact enclosure method."
],
"acceptance_gate": "Demonstrate independence, source binding, positive cases and semantic failure controls. Name the existing checks and the missing coverage. More digits or another equivalent evaluation of an unchanged root is not by itself new verification coverage.",
"first_step": {
"command": [
"python3",
"-B",
"scripts/research_baseline.py"
],
"expected": "PASS_RESEARCH_BASELINE_BINDINGS_ONLY",
"inspect": "Read #2979 joint rank-7 end-frame proposal and the revision-6 evidence status. Its claimed kappa is not selected. Compare coordinated changes with the losing single-change controls and locate the complete physical ledger; an aggregate moment alone does not validate the mechanism.",
"scope": "Positive control / source-bound orientation only; this is not a new construction or full cold replay."
},
"first_experiment": "Prioritize a bounded independent audit of #2979: reconstruct one joint group and its single-lowering control, check the declared frames/chronology, then trace changed children, idle widths and cleanup buffers into the complete invoice and both supplier caps. Choose one still-missing implication after checking existing audits. Stop with exact unresolved dependencies if full admission exceeds the session; do not launch a new lowering census on an unvalidated premise or substitute extra root digits for mechanism checking.",
"bottleneck_link": "Independent pricing prevents unsupported frontier claims and can reveal genuine accounting slack.",
"allowed_variations": [
"Reconstruct either a source ledger or a numeric enclosure with a different method."
],
"scope_change_triggers": [
"Importing the primary function and calling its output independent verification.",
"Promoting #2979 from merged source archive to validated supplier solely from author PASS receipts or aggregate pricing."
],
"overlap_terms": [
"independent moment",
"rank histogram",
"exact enclosure",
"2979",
"rank7 joint endframes",
"paired extraction",
"joint lowering"
]
},
"scope_status": "changed",
"actual_scope": "Independently audit PR #3647 (head 36ff4414)'s composed R437 bit supplier: the helper464 joint GL deletion composed with #1043's helper470 deletion, against single-change controls #589+E464 (the PR's R438) and #589+E470 (#1043's R438, rebuilt). Word level: the exact F2 identity on all 659 inputs with arbitrary dirty helpers, by a forward replay and a reverse pull-back, plus the literal reverse orientation; literal reconstruction of R437 from selected #589; aligned-trace non-interaction; SOURCE77 frame-186 chronology and all SOURCE write boundaries; the local profile and its deltas; exact rational projector identities as a second frame method. Stretch, completed: byte binding of the PR's records and the template's local programs; the parent schedule and idle windows from main's recipe and hosting rule; legality of the 3,933 slots per replica, and all 288 calls' owners and local-program hashes; the whole paid profile; the invoice precision envelope including the payload-prefix and ROW caps; and an independent exact price bracket for c, with positive controls reproducing selected #589's a and #1043's c. Plus 18 word mutants, 4 hosting mutants and 2 single-child price mutants, all rejected, and one positive control.",
"scope_change_reason": "The card's default object is a complex-moment/assembly claim (#2979). The loop's picker drew the bit-mechanisms track, and PR #3647 makes the highest open bit-supplier value (c = 0.001736628253427115). Apart from the author's own checks, it had only arithmetic or author-code checks: PR3668 checks the histogram-to-c arithmetic, and the concurrent PR3690 (opened during this audit) reconstructs paid requests with predecessor and author code. The card's method, reconstructing a coordinated change against its single-change controls, is applied to that supplier. Correction: this package's task selection was first justified by the claim that the bit side would bind if PR3653/PR3682 were admitted. That rested on a mixed-convention comparison (an 8.57e-9 gap between #589's retained cap a/(1+a) and #3653's modern-outer kappa) and is withdrawn. Under either consistent convention the bit side keeps about 3.0e-6 of headroom.",
"bottleneck_link": "Under the selected chain's inherited semantics, PR #3647's bit supplier c = 0.001736628253427115 (retained cap 0.0017336176; modern-outer bit value 0.0017366283) now has an independent finite check from the literal word through hosting to the price root. There is no near-term kappa consequence. The complex primitive binds today, and it would still bind if PR3653/PR3682 were admitted: the bit headroom is 2.998e-6 under the retained transfer and 3.008e-6 under the modern outer. The bit supplier caps kappa only after a further complex gain of about 3.0e-6, and this check then applies to the highest open bit value. It also records one item to confirm at intake, likely neutral: all nine captured retained totals of selected #589 already carry 42-66 helper-dirt columns at the COPY, and E464 only swaps which helper's column (the deleted h464's for h340's) reaches two of them, h477/h478, at unchanged counts 46/43.",
"budget_minutes": 180,
"stop_rule": "Stop at the first of: a confirmed discrepancy (minimize with mutants); a full PASS of identity, composition, controls, chronology, profile and mutants with both methods agreeing; word semantics not pinned within about 60 min (hand off); or a PR head change (re-pin once). Outcome: full PASS of the core within the semantics budget. The bounded hosted-invoice stretch (at most 60 min) also completed inside the experiment window: hosting, profile, precision envelope and price bracket. c is reproduced under the inherited semantics, not admitted.",
"outcome": "completed",
"evidence": "PASS_INDEPENDENT_R437_WORD_AUDIT (RESULT.json). Word: R437 (659 inputs) and controls #589 (661), #589+E464 = PR R438 (660) and #589+E470 = #1043 R438 (660, rebuilt) all map (x,y,h) to (x,y+x,h) over F2 with legal nested climbs and FULL/t-perp final frames. Forward replay and reverse pull-back agree entrywise; the reverse orientation passes; H1-H5/U/J hold; D0/C equal the rebuilt ones. #589+E464+E470 equals the PR's R437 except stats metadata. The E470 control reproduces #1043's FRAME-PATHS.json and its CERTIFICATE.json local profile, 24,248 ADDs and gates 111/2074/712/2573 (machine-checked), and D0 18,085->18,030 (computed; its equality with the value stated in PROOF.md is not machine-checked). The ten GL gates are taken as data from the PR's R438 word. E470 changes no projected value; E464 changes no clean column; with deleted columns projected away only column 340 differs, at A2044/A2047/A2068/A2070 and their C inverses. Capture census: #589's nine retained totals carry {42,44,46,63,66,46,43,62,59} helper-dirt columns at the scatter; in R437 h477/h478 swap h464's column for h340's (46/43 kept) and h475 loses h470's (63->62); nothing else changes. The composed trace equals the E464 control's with h470 deleted at all 5,369 aligned gates. Profile {1:1320,2:1344,3:77,4:43,5:102,6:11,7:18}, mass 5,113, ADDs 24,208; deltas additive. All 966 SOURCE write gates and all SOURCE climbs match #589; x77 is clean at the frame-186 read, reached at A426. Exact rational projectors: ranks equal dims for all 636 frames; P_A P_B = P_B P_A = P_A on all 2,384 climb pairs. Records: the PR stream (27,132 records, sha caeb2eb0...) is byte-identical to an independent lowering, replays over F2 in both orientations (norms 49,347/1,113,998; signs bound by the byte equality), and equals the template's local programs; all 288 calls' local_hash values equal the programs' SHA256. Hosting: the schedule equals main's recipe; the windows are rederived; 3,933 slots per replica are legal and disjoint; remainders give idle {2:431,4:444,6:444,7:666,34:2,42:11}; 288 calls' owners match the slots and data routes; 125,856 hosted. Whole profile {1:380160,2:400864,3:22176,4:26592,5:29376,6:17376,7:26496,34:64,42:352}, mass 1,808,352, deficit 10,272. Precision: prefix 177 (192), coefficient 177, two-level 393 (512), depth 2, ROW 18,355 < inherited cap 20,161, payload prefix 93 bits < 104. Price (m = 64, W = 28,416 read from the template and asserted): c = 868314126713557740143/(5e23) brackets with the next 1e-24 point failing. Controls: selected #589 a = 0.001733427039589844234749 from main-only inputs and #1043 c = 0.001734176288479865766692 are reproduced. Single-child price mutants fail the expected bound (2->3 at c; 42->34 at the next grid point). The early-read mutant passes F2 identity and legality and is rejected only by the chronology comparison, which the audit adopts as a sufficient inherited condition.",
"author": "Sagar Thakrar (thakrarsagar)",
"ai_assistance": "Substantial AI assistance from Claude (Anthropic, Claude Opus 5.5): design, code, checks and write-up in an automated research loop. Reviewed by seven independent AI review passes in the same loop, in two rounds (four in round 1, three in round 2); findings are addressed, and two suggested texts were not adopted (a replacement sentence repeating the mixed-convention comparison, and keeping a 32-child price mutant). No human or formal-kernel review is claimed.",
"open_questions": [
{
"question": "Does the inherited capture contract depend on which helpers' dirt reaches a captured total? Selected #589's nine retained totals already carry 42-66 helper-dirt columns at the COPY; in two of them (h477/h478) E464 swaps the deleted h464's column for h340's at unchanged counts 46/43, with clean rows unchanged.",
"why_it_matters": "The rank-7 captures rely on the inherited 'one paid transform plus linear private-stream discard' contract, which #589 already needs for dirty captures. If it is linear in arbitrary helper contents, as PR #3647 states, the item closes; only if it named particular helpers would the GL edit need a separate native argument.",
"first_step": "State the native capture lemma's hypotheses as used by selected #589 and check whether any of them names particular helpers.",
"blocked_by": "Locating the native capture lemma statement in the selected sources."
},
{
"question": "Does the E464 pattern recur: which other carriers of #589/R437 have a clean contribution recoverable from the output roles of the single node that reads them?",
"why_it_matters": "E464 is the new mechanism in PR #3647. A census would show whether it is isolated or a family, which bears on whether further bit-supplier gains of this kind exist.",
"first_step": "Run a read-only census of carriers with one birth whose reads are confined to one node frame, and record the five-point expansion decision before any construction round.",
"blocked_by": "The expansion gate; no construction is authorized by this audit."
},
{
"question": "Which admission dependencies remain for PR #3647 as a bit supplier?",
"why_it_matters": "A checked local supplier is not an admitted one. This is not urgent: no known complex claim makes the bit side bind (about 3.0e-6 of headroom remains after PR3653/PR3682).",
"first_step": "In the intake review of #3647, #3668 and #3690, map this package's checks and open question 1 against the inherited interface list.",
"blocked_by": "Maintainer intake of PR #3647."
},
{
"question": "Is PR #3647's hosting price-optimal under the audited rule? Merging the two stage-6 leftovers (helper indices 421 and 422, registers 643 and 644) into copy 209 of window 6 gives remainders {26:1, 42:12} instead of {34:2, 42:11} at the same mass, and raises the root by about 1.78e-8 (to 0.0017366460420).",
"why_it_matters": "It is a small, mechanism-free gain for the PR's author. It was found by one AI review pass and re-checked locally (not shipped) for slot legality and the price equation only.",
"first_step": "Regenerate the call owners for the merged slot, rerun the owners check and the physical-layer checks, then reprice.",
"blocked_by": "Owners and the physical layer for the merged slot are not checked."
}
]
}
Loading