docs: record h2 order-control scout results#312
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates the documentation and artifact indices to reflect the results of the H2 output-cloud geometry order-control scout. The changes confirm that the research-side signal remains strong after addressing the class-ordered seed-offset caveat. Feedback from the reviewer focuses on technical consistency within the new JSON artifact files, specifically recommending the use of forward slashes in file paths for cross-platform compatibility and correcting misleading or stale metadata fields.
| "track": "black-box", | ||
| "method": "H2 output-cloud geometry scorer", | ||
| "mode": "cpu-cache-review", | ||
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-class-ordered-subset-20260525\\response-cache.npz", |
There was a problem hiding this comment.
The path uses Windows-style backslashes. For cross-platform compatibility and consistency with the rest of the repository, please use forward slashes.
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-class-ordered-subset-20260525\\response-cache.npz", | |
| "response_cache": "workspaces/black-box/runs/h2-response-strength-256-class-ordered-subset-20260525/response-cache.npz", |
| "reopen_allowed": false, | ||
| "requires_reseeded_or_interleaved_cache_before_promotion": true | ||
| }, | ||
| "verdict": "weak_non_complementary_output_cloud_geometry", |
There was a problem hiding this comment.
The verdict weak_non_complementary_output_cloud_geometry is misleading given the high AUC (0.967) and strong strict-tail metrics. This likely occurs because the comparison block is null, but the label should reflect the actual signal strength.
| "verdict": "weak_non_complementary_output_cloud_geometry", | |
| "verdict": "candidate_output_cloud_geometry", |
| "track": "black-box", | ||
| "method": "H2 output-cloud geometry scorer", | ||
| "mode": "cpu-cache-review", | ||
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-class-ordered-subset-20260525\\response-cache.npz", |
There was a problem hiding this comment.
The path uses Windows-style backslashes. Please use forward slashes for consistency.
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-class-ordered-subset-20260525\\response-cache.npz", | |
| "response_cache": "workspaces/black-box/runs/h2-response-strength-256-class-ordered-subset-20260525/response-cache.npz", |
| "track": "black-box", | ||
| "method": "H2 output-cloud geometry scorer", | ||
| "mode": "cpu-cache-review", | ||
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-shared-position-20260525-r1\\response-cache.npz", |
There was a problem hiding this comment.
The path uses Windows-style backslashes. Please use forward slashes for consistency.
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-shared-position-20260525-r1\\response-cache.npz", | |
| "response_cache": "workspaces/black-box/runs/h2-response-strength-256-shared-position-20260525-r1/response-cache.npz", |
| "notes": [ | ||
| "This is a CPU-only scorer review on an existing H2 response cache.", | ||
| "It intentionally excludes seed-to-output distance features so it cannot collapse back into H2 simple distance.", | ||
| "A positive result is candidate-only until reseeded or interleaved response-cache controls rule out class-ordered sampling effects.", |
There was a problem hiding this comment.
This note is stale for this specific artifact. Since this file represents the shared-position order-control scout, it has already ruled out the class-ordered sampling effects mentioned in the note.
| "A positive result is candidate-only until reseeded or interleaved response-cache controls rule out class-ordered sampling effects.", | |
| "A positive result is candidate-only until formal promotion mechanisms or independent consumption contracts are established.", |
| "track": "black-box", | ||
| "method": "H2 output-cloud geometry scorer", | ||
| "mode": "cpu-cache-review", | ||
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-shared-position-20260525-r1\\response-cache.npz", |
There was a problem hiding this comment.
The path uses Windows-style backslashes. Please use forward slashes for consistency.
| "response_cache": "workspaces\\black-box\\runs\\h2-response-strength-256-shared-position-20260525-r1\\response-cache.npz", | |
| "response_cache": "workspaces/black-box/runs/h2-response-strength-256-shared-position-20260525-r1/response-cache.npz", |
1 participant
Summary
256 / 256shared-position order-control scout resultsDecision
512 / 512shared-position rerun selected by defaultVerification
python -X utf8 -m pytest tests/test_run_h2_response_strength_validation.py tests/test_review_h2_output_cloud_geometry_script.py -qpython -X utf8 scripts/check_markdown_links.pypython -X utf8 scripts/check_public_surface.pypython -X utf8 scripts/export_admitted_evidence_bundle.py --checkpython -X utf8 scripts/run_pr_checks.py