Skip to content

Deterministic Area grading (Memo 107) #57 - #58

Merged
a6b8 merged 1 commit into
mainfrom
memo-107-deterministic-area-grading
Jun 4, 2026
Merged

Deterministic Area grading (Memo 107) #57#58
a6b8 merged 1 commit into
mainfrom
memo-107-deterministic-area-grading

Conversation

@a6b8

@a6b8 a6b8 commented Jun 4, 2026

Copy link
Copy Markdown
Contributor

Closes #57.

Adds the deterministic Answer-Mapper + wiring so deterministic grading writes real `_gradings/` entries (not just a pretest summary).

  • `DeterministicAreaMapper` (script grader, no LLM): single-test (working-test bar) + tools-aggregate-schema (structural validity). gradingMode partial, tier-cap B, key-gated tools skipped.
  • `AreaScorer` re-exported from the entry.
  • DataPretest: response-size dimension (`responseBytes`/`large`/`extreme`) + reversible stale-test cleanup (`.trash/`, dryRun-safe).

Tests: 791 passing, lint:hygiene green. Version 2.3.0.

🤖 Generated with Claude Code

DeterministicAreaMapper (script grader, no LLM) + AreaScorer
re-export; response-size dimension + reversible stale-test
cleanup in DataPretest. v2.3.0.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@a6b8
a6b8 merged commit 78b9c08 into main Jun 4, 2026
1 check passed
@a6b8
a6b8 deleted the memo-107-deterministic-area-grading branch June 4, 2026 18:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Deterministic Area grading: DeterministicAreaMapper + stale-test cleanup

1 participant