Train and Validation (14 each) are Japanese passages written in Japanese, ModelWritten
Anthropic, generator claude-sonnet-5-5; none sits in a Test group. Test is eight
one-sentence excerpts (Soseki, Kyoka, Akutagawa, Ogai, Arishima, Kenji, Kajii, Kunikida)
taken from the card ruby zips (cp932, ruby and the pipe marker dropped, NFC) and checked
verbatim against each card's XHTML file; licence strings keep 底本, 入力 and 校正 within
the store's 200-byte cap. Entities appended (keeper gains 番人; eight pd-* authors).
Items 1..28 and holdout:1..8 unchanged (digest pinned). Tests per language: counts,
provenance, Test public-domain only, no markup, panel round reads both languages, speaker
forms, recall 56, no Test text in any prompt.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Eight one-sentence English excerpts from Dhammapada (Muller), Walden, Emerson,
Bacon, Aesop (Townsend), Mill, Plato (Jowett), Darwin, each taken from the
exact Project Gutenberg page named in edition_url and checked verbatim against
the fetched text. Groups hash into Test (buckets 83-97); split.rs untouched, the
ModelWritten-never-in-Test rule stands. Items 1..28 are byte-identical (digest
pinned). Tests: counts Train/Validation 14/14, Test 8 all PublicDomain; no Test
text in any round-1 or recall prompt or RoundRecord over data/original; a
ModelWritten item in a Test group is read in no split.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Six short original sentences, model-written (Anthropic, claude-sonnet-5-5) and so
labelled, three Train and three Validation, over the Dao revision-3 inventory and
one live entity. The recall probe now covers every offered item, naming the work
only for a public-domain one, so it can show the original set reads near the
guessing floor.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
