Certify Build O US Populace as the certified default#483
Merged
Conversation
BUILD_O_RELEASE_ID populace-us-2024-buildo-sparse-rmloss100-22bd902- 20260722T232627Z replaces populace-us-2024-buildn-sparse-rmloss100- c3e378a-20260722T010408Z; policyengine-us stays 1.764.6 (built_with_model_package), so pyproject and household snapshots are untouched - the #470 common case. Build O attempt 1 vs Build N: final loss 0.028131 vs 0.028242, within-10% 89.56% vs 89.49%, 5,672 targets on the same truthful surface (directly comparable, unlike N-vs-M). First release to pass every gate on its first attempt - the populace#491 register alignment at work. Input coverage 163/163 now includes the populace#500 adult-care and SE-health stages: CDCC adult-care reforms score real dollars instead of a silent $0. Known boundaries stated in full in the populace#368 verdict, led by the inherited populace#507 SSI aged-band defect (do not cite for SSI baselines); attempt 2 with the populace#508 fix supersedes this default when it lands. Strict-dominance rationale in the verdict. The UK TRO sidecar's sha fields update because it binds the same (changed) bundle manifest; not a degraded regeneration. Merging this PR is the default-flip decision and stays with the maintainer. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Merging this PR is the default flip to
populace-us-2024-buildo-sparse-rmloss100-22bd902-20260722T232627Z(published immutable 2026-07-22T23:26Z; latest.json untouched until merge), replacing Build N. Same #470 common-case shape as #475/#481: manifest rebind (sha5226bf5e…), both TROs regenerated with the HF token, three pinned test constants, towncrier fragment;policyengine-us==1.764.6unchanged.Why certify attempt 1 while attempt 2 is pending (owner rationale, full verdict on populace#368): strict dominance over the current default — loss 0.028131 vs 0.028242, within-10% 89.56% vs 89.49% on the SAME truthful surface, first-attempt gate-green, and input coverage 163/163 now including the populace#500 adult-care + SE-health stages (CDCC adult-care reforms score real dollars; on N they score $0). Its defect set is IDENTICAL to N's already-certified, already-documented set — led by the populace#507 SSI aged-band collapse, stated prominently in the verdict with the populace#508 fix gating attempt 2, which supersedes this default when it lands.
Verification: 738/738 pe.py battery green (4:11) against the published tag;
scripts/bundle.py checkrc 0; certify script run against the revision-pinned manifest URI;uv.lockuntouched.🤖 Generated with Claude Code