Skip to content

bench: publish assistant-run exploratory comparison #422

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Helweg merged 2 commits into from
Oct 6, 2026
Merged

bench: publish assistant-run exploratory comparison #422

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
976ff12
bench: preregister self-curated current-system comparison
Helweg Oct 6, 2026
94636cc
bench: publish assistant-run exploratory comparison results
Helweg Oct 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0

- **Competitive evaluation diagnostics**: Report per-task wins, losses, ties, metric deltas, ranked paths, operational failures and unsupported interfaces. Publish a dated historical replay and fresh-study protocol without treating development data as a blind holdout or claiming SOTA.
- **Study preparation evidence**: Publish a sanitized maintainer exposure register, current-version public-interface calibration and isolated broker/evaluator feasibility observations. Distinguish passing infrastructure controls from the failed bounded model repair attempt.
- **Assistant-run exploratory comparison**: Publish frozen nine-repository, 100-task results across seven conditions, all 2,100 query-pass records and a verified evidence replay. Report OCBI's exact-symbol and natural-language deficits, source-policy/declaration coverage failures and explicit self-curation limits; no outside-review prerequisite or SOTA claim.

### Fixed

Expand Down
Binary file added benchmarks/results/self-curated-2026-10-06/preregistration.json.gz
View file
Open in desktop
Binary file not shown.
Loading
Loading