populace-us-2024-9f1260b-20260611: eCPS-free build provenance - #6
Merged
Conversation
Primary-source-only construction (CPS ASEC, IRS PUF, Fed SCF 2022, SIPP, CPS-ORG, MEPS-IC, ACS 2022); enhanced CPS benchmark-only. Gates: parity 0, exported-nonzero pass (309 cols), calibration 95.09% within 10%, smoke in band (STCG -$77.5B vs -$76.8B signed target). Score (matched 41,314, symmetric refit): train 0.190 vs 1.089, holdout 0.038 vs 0.317, full 0.228 vs 1.406. Card + manifest + chain snapshots.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Release provenance for the fully primary-source rebuild: dataset card (de-eCPS'd construction + gates + score), release manifest (sha256 + revision pins), and verbatim snapshots of the chain that produced HF revision
4a8e7d39eb9e.Headline: parity 0 gaps, all 309 stored columns nonzero, calibration 95.09% within 10%, and the sound comparison at matched 41,314 households — train 0.190 vs 1.089, holdout 0.038 vs 0.317, full 0.228 vs 1.406 (per-target: incumbent 2,613 / populace 1,040 / 51 ties, disclosed on the card).
🤖 Generated with Claude Code