HCAI-Lab/w2-consensus-deepdive-unlearning-artifacts / social-data-attribution-w2 /src /data_attribution /evaluation
3.13 GB
11,407 files
Updated 8 days ago
Ctrl+K
| Name | Size | Uploaded | Xet hash |
|---|---|---|---|
| __pycache__ | 4 items | ||
| README.md | 1.11 kB xet | f8e08f9b | |
| __init__.py | 0 Bytes xet | 00000000 | |
| olmobaseeval.py | 2.12 kB xet | 689d002a | |
| olmobaseeval_registry.py | 5.16 kB xet | e2a79a62 | |
| olmobaseeval_subjects.py | 1.7 kB xet | 6bf8ede1 | |
| query_metadata.py | 3.29 kB xet | ebbb1637 | |
| socialiqa_classifier.py | 4.99 kB xet | 8fbcc41e | |
| socialiqa_patterns.py | 5.85 kB xet | f7dc7133 | |
| socialiqa_verified_labels.py | 7.05 kB xet | 456a8eca | |
| task_suites.py | 3.24 kB xet | 5dc07a64 |
Evaluation Package Scope
OlmoBaseEval
I keep OlmoBaseEval support as a separate expansion track from the current SocialTDA paper benchmark set.
The code in this package may live with the shared evaluation infrastructure
because it reuses OLMES export, query-manifest conversion, and TrackStar
analysis paths. That does not make base_olmobaseeval result artifacts part of
the SocialTDA paper outputs.
Current merge scope:
- Registry and conversion support for cloze-compatible OlmoBaseEval probes.
- Tests that prevent accidental export or aggregation of unintended probes.
- No tracked cluster manifests, rendered result pages, or paper-facing OlmoBaseEval result tables.
Result policy:
- Treat prior
base_olmobaseevalPACE and WIP outputs as separate-analysis artifacts. - Do not mix OlmoBaseEval results into SocialTDA paper result folders, report tables, figures, or claims unless a later release explicitly promotes that scope.
- Keep cluster-specific paths, run manifests, and large generated outputs out of the implementation PR unless they are reviewed as a separate results PR.
- Total size
- 3.13 GB
- Files
- 11,407
- Last updated
- Jul 1
- Pre-warmed CDN
- US EU US EU
