alea-institute/alea-count-length

ALEA count and segment

Counting and segmentation questions over real public-domain text and dictionary headwords: letters (with a spelled-out trace), occurrences of a letter, characters, UTF-8 bytes, words, non-blank lines, paragraphs and sentences, sentence segmentation in several output styles, and length-range checks.

119,699 documents, 12,143,734 tokens (alea-v2b-sw32k, the ALEA R4 tokenizer, counting each document's start and end tokens).

Part of the ALEA R4 training mix

This dataset contains exactly the documents of this family that ALEA's R4 model trained on in its stable pretraining mixture (second data stage, one pass each). Every row is byte-identical to the trained text; text_sha256 is the SHA-256 of text.

Held-out evaluation items generated by the same code are not part of this release.

Composition

task	documents	tokens
letters	21,619	743,490
letter_occ	17,748	767,901
words	17,206	2,215,741
lines	12,442	2,036,031
bytes	11,973	607,914
chars	11,681	568,388
sentences	7,672	890,702
seg_lines	5,704	1,140,677
length_range	5,420	1,084,379
seg_mark	4,137	789,131
paragraphs	4,097	1,299,380

41,998 documents use ALEA's chat template.

How it was made

Written entirely by deterministic Python generators; no language model wrote any of it. Every answer is computed in code from the passage, under a definition stated in the prompt (words are whitespace tokens, characters are code points, bytes are UTF-8, lines are non-blank lines, paragraphs are blank-line blocks). Sentence boundaries come only from the nupunkt sentence splitter, and a passage is used for sentence items only when nupunkt's two modes agree with high confidence. Hard-wrapped renders use textwrap without hyphenation. Prompts always put the instruction first; presentations are plain, Q/A and chat, with answers given bare, as a sentence, or with a worked trace.

Sources

• Federal Register daily XML paragraphs (2003–2023), eCFR, U.S. Code, GovInfo, Public Laws and GAO passages (U.S. government works, public domain)
• OpenGloss v2.4 headwords (mjbommar/opengloss-v2.4-lexicon, CC-BY-4.0)

Verification

• Every count is recomputed from the passage in code.
• Segmentation answers are re-verified to reproduce the source passage exactly once their markers are stripped.
• Paragraphs with e-mail addresses or phone numbers were dropped (detection only; source text is never rewritten).
• Decontamination: training rows that shared a word 13-gram with a held-out evaluation item of this or any sibling family were removed, and rows overlapping ALEA's evaluation sets were removed or logged according to ALEA's decontamination policy.
• Sampled records were reviewed blind by a separate reviewer who recomputed or re-derived the answers; defect patterns found in review were fixed in the generator, or the affected rows dropped, before the final build.

Columns

column	content
id	row id from the generator (not always unique: a few families reuse an id across variants, and a handful of exact duplicate documents were trained twice and appear twice)
text	the document exactly as trained; chat documents use ALEA's chat template, whose control tokens appear as text (`<
task	task kind
phase	stable (trained in R4's stable pretraining mixture)
generator	what wrote the row (deterministic = code only)
group_id	source passage / record group; all variants of one group share it
prompt_chars	length in characters of the request part of text (0 for plain documents; null where not recorded)
n_tokens	tokens under alea-v2b-sw32k, including the document start and end tokens
text_sha256	SHA-256 of text
meta	JSON: the generator's structured fields (source, presentation, answer form, gold values, checks)

Known limitations

• Sentence items cover only passages where nupunkt is confident; abbreviation-heavy and citation-heavy sentences are under-represented.
• The documents are generated and more repetitive than natural text.

License

CC-BY-4.0 for this compilation and the generated text. Public-domain and public-record source text keeps its status; Wikidata is CC0; OpenGloss and SHELF are CC-BY-4.0.

Published by the ALEA Institute.