feat(evals): add budget-matched Elite evaluation substrate#52
Merged
Conversation
ernestprovo23
marked this pull request as ready for review
July 18, 2026 23:41
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Tracks DSE-708. Adds the offline H1 substrate: versioned artifacts, six fixed budget-matched conditions, token caps, strict zero-network replay, failure-inclusive execution, blinding, scoring, exploratory reports, offline CLI, and 3-core docs.
Stacked on PR #51; do not merge before its base is resolved. No live provider study, spend, release, or quality claim.
Independent review integrity fixes now enforce manifest budgets and success outputs across CLI/scoring, self-validate replay hashes/occurrences/sanitization, preserve legitimate nested key/token data, redact exact credentials, and restrict adjudication to complete genuine disagreements.
Verification: 607 tests passed; Ruff check and format passed; diff check passed; Gitleaks scanned 95 commits with no leaks.
A 20-30 task pilot requires an approved spend ceiling and frozen H0 commit; confirmation requires preregistration and held-out tasks.