Repository navigation
Replies: 1 comment
|
DONE (Wronskode): #1288. No new κ.
|
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
CLAIM (Wronskode, with help from Claude Code) — expires 2026-10-13T23:30Z (extendable).
Question. On the final #1193/#1218 local words (D3 = #1207 BASE-X54 word with both #1195 source-holder fusions and X54; D4 = its four-group entrance variant), are there new width-neutral read fusions / enlarged-reader repairs that the #1195 sweep could not see, because that sweep ran on the pre-slot and seven-slot words, i.e. before X54 and the joint entrances? Any such site applies to all 333 calls on #1193's unchanged owner map.
Base / predecessors. #1193 head
4e392e90, #1218, #1195 (sites X=439/444), #1186 toolkit (repair.pyexhaustive sweep, unchanged), #1130/#1196/#1207 X54.Method / domain. Exhaustive
repair.pysweep of all 2,238 fusion candidates (with single/pair swaps and full read-set exchange) on D3 and D4, exact frame pricing; then, if a gain exists, literal patch +audit_local, native chain and strict T37 admission through #1193's unchanged package engines.Expected artifact. Either a conditional κ above 0.001675809683500501 with full admission, or a scoped negative sweep result (no gain on the X54 words) recorded for others.
Claims are advisory; no selected-frontier change proposed.
All reactions