Shape ledger: a semantic miss is not evidence; floors sized for the reader (#4208, #4306) #182

Merged

bvandeusen merged 3 commits from dev into main 2026-09-22 09:50:34 -04:00

Conversation 0 Commits 3 Files Changed 4

+186 -347

bvandeusen commented 2026-09-22 09:50:26 -04:00

Owner

Copy Link

Copy Source

• a875a1b: the divergence check no longer reads a semantic miss as "not the canon". Measured live, true instances rarely clear the arm's floor either, so the miss was silencing real divergences along with helpers (#4208).
• 312dc9f: the semantic proposer now uses the write-path floor, 0.68, instead of its own private 0.8. At 0.8 it proposed 0 of 150 while true instances scored 0.68–0.71. It now scans 8 matches instead of 3, and the proposer version bumps to 5 so rows are re-read.
• a01deeb: the signature floor moves from 0.8 to 0.75 so model classes with a different mixin list (~0.795) get proposed (#4306).

CI is green on all three (runs 7258, 7262, 7266).

🤖 Generated with Claude Code

https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

- **a875a1b**: the divergence check no longer reads a semantic miss as "not the canon". Measured live, true instances rarely clear the arm's floor either, so the miss was silencing real divergences along with helpers (#4208). - **312dc9f**: the semantic proposer now uses the write-path floor, 0.68, instead of its own private 0.8. At 0.8 it proposed 0 of 150 while true instances scored 0.68–0.71. It now scans 8 matches instead of 3, and the proposer version bumps to 5 so rows are re-read. - **a01deeb**: the signature floor moves from 0.8 to 0.75 so model classes with a different mixin list (~0.795) get proposed (#4306). CI is green on all three (runs 7258, 7262, 7266). 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

bvandeusen added 3 commits 2026-09-22 09:50:26 -04:00

fix(shapes): a semantic miss is not evidence — the divergence check stops reading it (#4208) ... a875a1b2ee

Measured live: true instances of the service-unit canon score 0.68-0.71
against it at best, helpers 0.66-0.75 against unrelated snippets, and
nothing reaches the 0.8 floor. The "conclusive miss" fired for nearly every
body and silenced real divergences exactly as it silenced helpers. The
stored miss basis, its flag withdrawal and the report plumbing are removed;
the floor stays at 0.8 and the new-shapes-first ordering stays. False
prompts are answered by judgment (exempt with a reason).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

fix(shapes): the semantic arm proposes at the write-path floor, not a private 0.8 (#4208) ... 312dc9f6f7

0.8 was sized for proposals nobody reads; every semantic proposal is read
by a judge before it is confirmed. Measured, it proposed 0 of 150 while
judged instances of a canon score 0.68-0.71 - the write-path hint's own
floor asks the same question of the same documents at 0.68. The arm now
uses that floor, scans 8 hits instead of 3 (a true instance ranked 4th
behind snippets of other language families), and the proposer version
bumps to 5 so rows examined under the old floor are read again.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

fix(shapes): the signature basis admits the band true instances score in (#4306) ... a01deeb851

Measured over 459 judged instance rows against every same-family canon:
true instances score 1.0 or ~0.795 (a model whose mixin list differs from
the canon's), and 0.8 cut the second band entirely. The proposal is read
before it is confirmed, so the floor admits it: 0.75 takes every measured
true row, at 35 wrong-canon pairings instead of 14.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

bvandeusen merged commit bf2871ffb4 into main 2026-09-22 09:50:34 -04:00

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

No items

No labels

Milestone

No items

No Milestone

Projects

Clear projects

No projects

Assignees

Clear assignees

renovate-bot bvandeusen

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: bvandeusen/FabledScribe#182