Add eval/: WGU extraction baseline vs 22-field CTDL schema#270
Draft
manueljhc wants to merge 1 commit into
Draft
Add eval/: WGU extraction baseline vs 22-field CTDL schema#270manueljhc wants to merge 1 commit into
manueljhc wants to merge 1 commit into
Conversation
Evaluation of the current Credential extractor against 5 WGU program pages, scored field-by-field against on-page content and WGU's published Registry records. Baseline 4-field schema reaches 8.3% recall of on-page CTDL fields; a 22-field schema on the same input reaches 90.5%. - REPORT.md / REPORT.html: results, method, ranked improvement proposal - harness/: standalone Node runner reproducing baseline + enhanced runs - wgu/<slug>/: source HTML, model input, registry ground truth, outputs - scoring/: 90-cell field matrix + per-page and per-field scorecards - code-review/: architecture notes + 10 findings with file:line refs
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Evaluation of the current Credential extractor against 5 WGU program pages, scored field-by-field against on-page content and WGU's own published Registry records.
Headline: the baseline 4-field schema captures 8.3% of on-page CTDL fields; a 22-field CTDL-aligned schema on the same simplified-markdown input reaches 90.5%. The gap is almost entirely schema/prompt coverage, not model capability.
Start here →
eval/REPORT.html(orREPORT.md) — results, method, ranked improvement proposal, and a design for a repeatable eval, all in one doc.What's in
eval/REPORT.html/.mdharness/run.mjs,xtra-prompts.mjs(verbatim baseline fromcommon/catalogueTypes.ts),enhanced-schema.mjs(proposed 22-field),simplify.mjswgu/<slug>/page.html·page-annotated.md·model-input.md·registry.json·extracted.jsonscoring/code-review/file:linereferences intoserver/andcommon/ctdl-reference/field-spec.mdReproduce
Opening as a draft for discussion — happy to reshape where this lands in the repo, or split the harness from the report if that's a better fit.