Datasets:
NuBerea/literary-analogy
An access-controlled registry of attested literary analogies and narrative parallels in the Hebrew Bible. It represents parallel loci, their passage members and spans, structured scholarly attestations, and an explicitly provisional verse-alignment surface.
The registered NuBerea tools expose four governed query views: passage profiles,
pair alignments, parallel-family browsing, and attestation provenance. Internal
bookkeeping and annotation configs are not exposed as agent tools. The complete
schemas, methodology, row counts, and provenance specification are contained in
the gated SPEC.md accompanying this dataset.
Reproducibility floor (2026-08-17)
The eight la_* registry configs are deterministic merges. The tables they merge that
are not themselves rebuildable from Hugging Face are published alongside them as tier-0
inputs so the registry is reproducible from this repository plus its declared upstreams:
hand-curated seed registry (la_seed_*, la_hard_negatives), the assembled event-chain
extraction (la_event_chains, la_event_chains_ambiguity), fact-grade scholarship claim
rows (la_research_sweep_claims, la_scholarship_prose_claims; references, locators and
claim types only — no source prose), and frozen model outputs (la_miqrabert_verse_vectors,
la_abstracts, la_abstract_vectors).
License and attribution
This dataset contains structured, fact-grade assertions and identifiers. It does not redistribute quoted scholarly prose or extrabiblical source text. Bibliographic attributions remain attached to the relevant assertion rows. Access is manually gated under the NuBerea estate terms recorded in the full specification.
- Downloads last month
- 52