fix: Cat9b None-vs-0.0, flat_rarity_mode flag, spec caveats, starter corpus, v0.2.0#28
Open
jphein wants to merge 1 commit into
Open
fix: Cat9b None-vs-0.0, flat_rarity_mode flag, spec caveats, starter corpus, v0.2.0#28jphein wants to merge 1 commit into
jphein wants to merge 1 commit into
Conversation
…tarter corpus, v0.2.0 Ports five fixes from #14 (fork issues #5-10) that were not yet PR'd upstream: - Cat 9b call_through_rate returns None (not 0.0) when the adapter declares no harness manifest — distinguishes "not measured" from "every probe failed" (#7) - gap_detection surfaces flat_rarity_mode flag when the rarity fallback fires on ≤2-sized components, so consumers know scores are in the flat regime (#8) - sme_spec_v8.md adds status caveat and 🚧 markers for unimplemented CLI commands (snapshot, run, compare, calibrate, viz, report) (#6) - standard_v0_1/questions.yaml + AUTHORING.md — reference questions and authoring guide for the starter corpus (#9) - Version bump 0.0.1 → 0.2.0 (#10, #5) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Ports five fixes from
techempower-org/multipass-structural-memory-eval#14(fork issues #5-10) that shipped on the fork but were never PR'd upstream. All are small, self-contained, and independently testable.Nonevs0.0semantic fix —call_through_ratenow returnsNonewhen the adapter declares no harness manifest, distinguishing "not measured" from "every probe failed" (techempower-org/multipass-structural-memory-eval#7)flat_rarity_modeflag on gap detection — surfaces when the rarity fallback fires on ≤2-sized components, so consumers know scores are in the flat (misleading) regime (techempower-org/multipass-structural-memory-eval#8)techempower-org/multipass-structural-memory-eval#6)standard_v0_1/questions.yaml+AUTHORING.md— reference questions and authoring guide for the starter corpus (techempower-org/multipass-structural-memory-eval#9)0.0.1→0.2.0— reflects the actual feature surface after PRs feat(cat9): minimum-viable Category 9 scaffolding + 9b call-through success #1-feat(adapters): RlmAdapter + Qwen-7B/Llama-70B baselines #7 (techempower-org/multipass-structural-memory-eval#10,techempower-org/multipass-structural-memory-eval#5)Test plan
pytest tests/test_gap_detection.py tests/test_harness_integration.py -v— 22 passed, 1 skipped (Betti-1 optional dep)ruff checkon all changed files — cleanflat_rarity_modetests: flagging, no-candidates, format-report texttest_cat9b_empty_manifest_reports_not_applicable— verifiesNonereturn🫏 Generated with Claude Code