-
card: brand GPTKB 2.0 (matches the papers)
-
card: drop em-dashes from prose
-
rename full record to gptkb_v2.0.1_full_record (align with the RDF cut); cite published papers
-
dataset card: v2.0.1 RDF + full-record Parquet release
-
v2.0 full record: six-term facts + catalogs + batch provenance (Parquet)
-
v2.0.1 RDF re-export: disambiguation complete (NE coverage 99.8%)
-
Add dataset card: license, vocabulary, and what the RDF view drops
-
Add GPTKB v2.0 simplified RDF export (N-Triples, gzipped)
-
initial commit
