Repository navigation
Run the declared held-out test evaluation of the CRF and the run-3 DeBERTa - #61
Merged
Merged
Conversation
…BERTa Declared by the owner on 2026-11-15: "Final held-out evaluation of v0.1.0 (CRF) and the run-3 DeBERTa candidate; no hand labels will come." The frozen systems (rules-2, crf-batches-01-02 from v0.1.0, the run-3 DeBERTa from model-deberta-3-run3) were scored once on the locked test split; every run is in docs/test-access-log.md. On the 146 model-drafted test records (agreement, not accuracy) the CRF leads: typed F1 0.193 against 0.154 for DeBERTa (paired -0.039 [-0.080, -0.000], clear only just) and 0.048 for rules. The CRF is clearly more precise and invents clearly fewer prerequisites; DeBERTa has clearly higher recall and fewer reversed meanings. On the 4 hand-labelled test records no difference is clear, so the release gates cannot be judged; no target changed.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Declared release evaluation (owner, 2026-11-15): "Final held-out evaluation of v0.1.0 (CRF) and the run-3 DeBERTa candidate; no hand labels will come." Run once; all 10 scoring runs are appended to
docs/test-access-log.md.rules-2, CRFcrf-batches-01-02(v0.1.0), DeBERTa run 3 (downloaded frommodel-deberta-3-run3, SHA-256 checked). Before scoring, each system's predictions on the 88 non-test pilot documents were confirmed identical to its committed dev predictions.scripts/experiments/test_evaluation.sh "<reason>"; results and per-class support inreports/test-evaluation/.DeBERTa − CRF typed F1 −0.039 [−0.080, −0.000] (clear, only just). 4 hand-labelled test records: no clear difference; release gates not judgeable (no target changed). README, model card, dataset card, evaluation doc, targets changelog and annotations README updated.