Of the 1,811 citations in Titles 34, 36 and 37, 1,519 (84%) resolve to
a section in this dataset; the card names the most-cited outside titles
(68, 49, 33, 8) and counts the rest. Computed from crossrefs at export
time. Data files are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add the embeddings subset; fix section reassembly
New `embeddings` subset: one openai/text-embedding-3-large vector per
chunk (3,072 float32, unit length), with the exact string embedded and
its SHA-256, exported from the author's index after it verified against
the corpus. The card documents the model, provider pin, the embedded
string, a search example and the measured weak-match score.
Fix: a section's chunks rejoin with "\n", not "\n\n"; that reproduces
all 4,178 sections exactly. Chunk and edge files are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A five-step pattern for agents (find, read the whole section, check the
version in force, follow references both ways, cite precisely) with
runnable helpers: read_section, versions, in_force and references.
Data files are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Explain that chunk text is Markdown (cite and heading as headings, bold
full label paths) and show the opening of tn:37-1-101:0001, taken from
the data at export time. Data files are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
9,050 chunks and 6,127 cross-reference edges as Parquet, with the
dataset card. Current through the 2026 Regular Session and the 2026 2nd
Extraordinary Session (publisher datestamp 2026-09-30).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
