[
 {
  "id": "TLL-01",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the record's name for the entity",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].name",
  "quote": "The Pile",
  "kernel": false,
  "const": false,
  "lineage": "what it is: 22 datasets combined",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (name)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-02",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its other names: the Pile, Pile",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].aka",
  "quote": "\"the Pile\", \"Pile\"",
  "kernel": false,
  "const": false,
  "lineage": "what it is: 22 datasets combined",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (aka)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-03",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its operator: EleutherAI, from the footer and the paper's author line",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].operator",
  "quote": "EleutherAI (the home page footer reads 'EleutherAI 2021'; the paper's authors are Leo Gao and 11 others)",
  "kernel": false,
  "const": false,
  "lineage": "who made it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (operator)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-04",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its kind: an open source language modelling data set, an English corpus for training large models",
  "kind": "definition",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].kind",
  "quote": "open source language modelling data set: an English text corpus targeted at training large-scale language models",
  "kernel": false,
  "const": false,
  "lineage": "what it is: 22 datasets combined",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (kind)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-05",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its area: the corpora",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].areas",
  "quote": "corpora",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (areas)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-06",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its evidence: dataset, documentation, paper",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].evidence_kind",
  "quote": "\"dataset\", \"dataset documentation\", \"research paper\"",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (evidence_kind)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-07",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what it is to the census: infrastructure",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].host_relation",
  "quote": "infrastructure",
  "kernel": true,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (host_relation)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-08",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its record type: dataset",
  "kind": "definition",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record_type",
  "quote": "dataset",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record_type)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-09",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the question it answers: training membership by component",
  "kind": "definition",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].question",
  "quote": "training membership by component: which 22 component datasets make up the training set, and which of them carries a given work?",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (question)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-10",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its unit: a component dataset and, within it, a document",
  "kind": "definition",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].unit",
  "quote": "a component dataset (one of 22) and, within it, a document; distributed as jsonlines compressed with zstandard",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (unit)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-11",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what it records: 22 subsets, existing and new, books, transcripts and papers among them",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record_object",
  "quote": "22 diverse subsets, both existing and newly constructed, many from academic or professional sources; per Wikipedia, books, movie transcripts and scientific papers among other media, with components including Pile-CC, PubMed Central, Books3, OpenWebText2, arXiv, GitHub, Gutenberg (PG-19), OpenSubtitles, Wikipedia, YouTube Subtitles and Enron Emails",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record_object)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-12",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "in the layer: the texts of its 22 components; Books3 from Bibliotik, removed; YouTube subtitles scraped",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record.in_layer",
  "quote": "the texts of its 22 components, assembled for training large language models; per Wikipedia, the Books3 component held copyrighted material compiled from Bibliotik, a piracy website, and was removed from the Pile; tens of thousands of YouTube videos had their subtitles scraped into it",
  "kernel": true,
  "const": false,
  "lineage": "Books3: source, takedown, suit",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record.in_layer)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-13",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "names: not recorded in its own saved pages; three authors sued over Books3",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record.name",
  "quote": "not recorded in its own saved pages; per Wikipedia, three authors filed a class action in 2024 over the Books3 component",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record.name)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-14",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "readback: not recorded; download and a benchmark, no per-work search; taken down from its site by 2024",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record.readback",
  "quote": "not recorded: its own pages offer download and a bits-per-byte benchmark (Pile BPB), with no per-work search; per Wikipedia, by 2024 it was taken down from its original site and remained accessible from other file sharing services",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record.readback)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-15",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what the apparatus makes: language models, scored on Pile BPB",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].record.makes",
  "quote": "language models trained on it: its own page scores models on Pile BPB, and the paper reports that models trained on the Pile improve over Raw CC and CC-100 on all its components",
  "kernel": true,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (record.makes)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-16",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its method: 22 datasets combined, code public, concerns documented",
  "kind": "function",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].method",
  "quote": "combination of 22 smaller datasets into one training set; code used in its construction made public; an exploratory analysis documenting 'potentially concerning aspects of the data'",
  "kernel": false,
  "const": false,
  "lineage": "what it is: 22 datasets combined",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (method)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-17",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its coverage: English text, released 31 December 2020",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].coverage",
  "quote": "\"geography\": \"English text\", \"period\": \"released 31 December 2020\"",
  "kernel": false,
  "const": false,
  "lineage": "who made it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (coverage)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-18",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its count: 825 GiB, 886.03 GB, 22 components",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].count.value",
  "quote": "825 GiB (own pages; titled 'An 800GB Dataset'); 886.03 GB (Wikipedia's component table); 22 component datasets",
  "kernel": false,
  "const": false,
  "lineage": "the size given three ways",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (count.value)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-19",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its holdings: the largest components by Wikipedia's table",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].holdings",
  "quote": "\"per Wikipedia, largest components Pile-CC (243.87 GB), Books3 (108.40 GB), GitHub (102.18 GB), PubMed Central (96.93 GB)\"",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (holdings)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-20",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its stated limits, among them the size given three ways",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].stated_limits",
  "quote": "\"size given three ways: '800GB' (title), '825 GiB' (home and abstract), 886.03 GB (Wikipedia)\"",
  "kernel": false,
  "const": false,
  "lineage": "the size given three ways",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (stated_limits)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-21",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what it answers: its components and sizes; a model's Pile BPB",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].answers",
  "quote": "\"which 22 component datasets make up the Pile and their sizes\", \"how a model scores on Pile BPB\"",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (answers)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-22",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what it does not answer: whether a given work is in it; under what author name",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].does_not_answer",
  "quote": "\"whether a given work is in the Pile (no per-work search in its own pages)\", \"under what author name a work sits in a component\"",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (does_not_answer)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-23",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "what it does not answer: where it is now hosted",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].does_not_answer.2",
  "quote": "where the Pile is now hosted (own page names the Eye; Wikipedia says it was taken down from its original site by 2024)",
  "kernel": false,
  "const": false,
  "lineage": "where it is held",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (does_not_answer.2)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-24",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the archive's reading: the training layer named and itemized; entry by component; what the apparatus makes is a model scored on the Pile's own benchmark",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].tll_reading",
  "quote": "The archive reads the Pile as the training layer named and itemized: a work enters by belonging to a component (Books3, PG-19, Pile-CC), and what the receiving apparatus makes from it is a trained model, scored on the Pile's own benchmark.",
  "kernel": true,
  "const": false,
  "lineage": "the training layer itemized",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (tll_reading)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-25",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the archive's reading: no per-work search; inspected from outside; the Books3 removal and the suit as the field's clearest case of works found in the layer under their authors' names after the fact",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].tll_reading",
  "quote": "Its own pages offer no per-work search; the corpus has since been inspected from outside (Data Portraits documents it), and Wikipedia's record of the Books3 removal and the authors' suit is the field's clearest case of works found in the layer under their authors' names after the fact.",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (tll_reading)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-26",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its access: download from the Eye, jsonlines in zstandard, code public; taken down by 2024",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].data_access",
  "quote": "download (hosted by the Eye, per its own page); jsonlines compressed with zstandard; construction code public; per Wikipedia, taken down from its original site by 2024",
  "kernel": false,
  "const": false,
  "lineage": "where it is held",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (data_access)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-27",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its licence: not stated in the saved sources",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].license",
  "quote": "not stated in the saved sources (arXiv 'view license' link unread)",
  "kernel": false,
  "const": false,
  "lineage": "licence and access",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (license)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-28",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its lifecycle: historical; Books3 removed; taken down by 2024; the Common Pile v0.1 in June 2025",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].lifecycle.detail",
  "quote": "released 31 December 2020; per Wikipedia, some datasets removed after 2023 copyright disputes, Books3 removed, the Pile taken down from its original site by 2024; EleutherAI and partners released Common Pile v0.1 in June 2025, containing only works whose licenses permit AI training",
  "kernel": false,
  "const": false,
  "lineage": "the successor: Common Pile v0.1",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (lifecycle.detail)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-29",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "its state: historical",
  "kind": "fact",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].lifecycle.state",
  "quote": "historical",
  "kernel": false,
  "const": false,
  "lineage": "the successor: Common Pile v0.1",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (lifecycle.state)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-30",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "negative tag: not the Common Pile v0.1",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].negative_tags.0",
  "quote": "NOT Common Pile v0.1, its 2025 successor of licensed works",
  "kernel": false,
  "const": false,
  "lineage": "the successor: Common Pile v0.1",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (negative_tags.0)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-31",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "negative tag: not searchable by work",
  "kind": "fact",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].negative_tags.1",
  "quote": "NOT searchable by work in its own pages",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (negative_tags.1)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-32",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "negative tag: not the same size in every source",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].negative_tags.2",
  "quote": "NOT the same size in every source: 800GB, 825 GiB and 886.03 GB are all given",
  "kernel": false,
  "const": false,
  "lineage": "the size given three ways",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (negative_tags.2)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-33",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "negative tag: not Common Crawl; Pile-CC one component, its derivation unstated in the saved sources",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].negative_tags.3",
  "quote": "NOT Common Crawl: Pile-CC is one of its 22 components, and the saved sources do not state its derivation",
  "kernel": false,
  "const": false,
  "lineage": "its standing among training corpora",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (negative_tags.3)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-34",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "relations: uses PG-19; neighbour of Common Crawl and of the Books3 search tool; used by Data Portraits",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].relations",
  "quote": "\"rel\": \"used_by\", \"target\": \"data-portraits\", \"note\": \"the Data Portraits paper documents the Pile with its data-sketch tools\"",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (relations)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-35",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "unchecked: the paper's full text, the licence page, the Eye, the suit, Pile-CC's derivation",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].unchecked",
  "quote": "\"the paper's PDF and HTML full text (only the arXiv abstract page is saved), so the datasheet and component documentation are not read\"",
  "kernel": false,
  "const": false,
  "lineage": "what it documents of itself",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (unchecked)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-36",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "founded: 2020, released 31 December 2020",
  "kind": "genealogy",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "census.json rows[slug == 'the-pile'].founded",
  "quote": "2020 (released 31 December 2020)",
  "kernel": false,
  "const": false,
  "lineage": "who made it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (founded)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (non-string values as their JSON text; whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-37",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the census's organizing question",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "census.json organizing_question",
  "quote": "How is literature addressed to a receiving apparatus, what does that apparatus make from it, and how can transmission, transformation and provenance be inspected across later reception?",
  "kernel": true,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (organizing_question)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (whitespace-normalised)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-38",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the census's admission rule",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "census.json admission_rule",
  "quote": "A row is an aggregator: a resource that gathers and holds many other works or records and says what it holds (operator, 2026-11-01).",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (admission_rule)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (whitespace-normalised)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-39",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the census's anchors, #186 and #1543",
  "kind": "fact",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "census.json anchor",
  "quote": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey … Generative Uptake: Training-Layer Literature Extended from Machine Reading to Machine Composition (EA-TLL-COMPOSITION-01)",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (anchor)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (whitespace-normalised)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "TLL-40",
  "dep": "TLL",
  "record": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile",
   "label": "TLL"
  },
  "claim": "the census's area 'corpora'",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "census.json areas.corpora",
  "quote": "The corpora: collections, crawls and datasets whose texts enter machine training",
  "kernel": false,
  "const": false,
  "lineage": "what the census asks of it",
  "date": "2026-11-01",
  "title": "traininglayerliterature.org layer census record: The Pile (https://www.traininglayerliterature.org/aggregators/the-pile), editor Lee Sharks, compiled 2026-11-01, schema 1.1",
  "creator": "Lee Sharks (editor)",
  "file": "/home/claude/fleet/traininglayerliterature-org/data/layer-census/census.json (areas.corpora)",
  "quote_ok": true,
  "quote_check": "within the field's value as loaded (whitespace-normalised)",
  "entity": "the-pile",
  "surface": {
   "url": "https://www.traininglayerliterature.org/aggregators/the-pile"
  },
  "note": "composed in session on 2026-11-01 from sources saved that day (data/layer-census/sources, sha256 in the manifest); its own pages and the arXiv abstract read, with the Wikipedia article and the Data Portraits abstract as third parties"
 },
 {
  "id": "CAP-1-01",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the components itemized; the size given both ways",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[0]",
  "quote": "THE LAYER ITEMIZED. The twenty-two components in the order of Wikipedia's size table, Pile-CC first and Enron Emails last, each with a gloss; the size given both ways, '825 GiB to 886 GB'",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-02",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the components named, no work and no author",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[0]",
  "quote": "The census row reads a work's entry into the layer as membership of a component; the composition names the components and no work and no author in any of them.",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-03",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the token count and model families on no saved source",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[0]",
  "quote": "'roughly 300 billion tokens' and the three model families are on no saved source; the Hugging Face card gives '100B<1T tokens'.",
  "kernel": false,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-04",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "Books3 glossed by genre",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[1]",
  "quote": "BOOKS3 GLOSSED. 'Books3 (Long-form literary text/books)'.",
  "kernel": false,
  "const": false,
  "lineage": "Books3: source, takedown, suit",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-05",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the source, takedown, removal and suit absent",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[1]",
  "quote": "The source, the Rights Alliance's DMCA takedown of July 2023, the removal and the three authors' suit are absent; the census row reads that suit as 'the field's clearest case of works found in the layer under their authors' names after the fact'.",
  "kernel": true,
  "const": false,
  "lineage": "Books3: source, takedown, suit",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-06",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "subtitles without their account",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[2]",
  "quote": "THE SUBTITLES AND THE SITE. 'subtitles' stands beside Books3 without its account",
  "kernel": false,
  "const": false,
  "lineage": "subtitles: OpenSubtitles and YouTube",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-07",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the takedown absent; the first card the original site",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[2]",
  "quote": "Wikipedia's 'By 2024, The Pile also was taken down from its original site' is absent, and the first card is the original site's home page",
  "kernel": false,
  "const": false,
  "lineage": "where it is held",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-08",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "models named; the benchmark not carried",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[3]",
  "quote": "The census row's 'makes' is the trained model scored on the Pile's own benchmark ('Pile BPB is a measure of world knowledge and reasoning ability in these domains'); the composition names models trained on it and does not carry the benchmark.",
  "kernel": false,
  "const": false,
  "lineage": "the benchmark: Pile BPB",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-09",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the paper's concerning aspects absent",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[3]",
  "quote": "The paper's 'we document potentially concerning aspects of the data for prospective users' is absent.",
  "kernel": false,
  "const": false,
  "lineage": "what it documents of itself",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-10",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the successor given without partners or date",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, findings[4]",
  "quote": "The composition gives EleutherAI alone and no date; the seventh card carries the date in EleutherAI's words",
  "kernel": false,
  "const": false,
  "lineage": "the successor: Common Pile v0.1",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-11",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the innocent reading first",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, reading",
  "quote": "The innocent reading first. Asked for the dataset, the Overview gives its maker, date, size, purpose and the models trained on it, lists its twenty-two components with a gloss on each, and closes on its copyright disputes and its licensed successor.",
  "kernel": false,
  "const": false,
  "lineage": "reception: the encyclopedia, the cards, the census",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-12",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the analysis: composed at its grain; the literary component glossed by genre; source, removal and suit reduced to content adjustments",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, analysis",
  "quote": "The dataset composed at its grain and itemized to its twenty-two components, the training layer named as the census row names it. Its literary component is glossed by genre, and the pirated source, the removal and the authors' suit that put works in it under their authors' names are reduced to 'content adjustments'.",
  "kernel": true,
  "const": false,
  "lineage": "reception: the encyclopedia, the cards, the census",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-13",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the record's title",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, mt",
  "quote": "'THE PILE DATASET' ON THE OVERVIEW — THE DATASET ITEMIZED, ALL TWENTY-TWO COMPONENTS; BOOKS3 AS 'LONG-FORM LITERARY TEXT', ITS PIRATED SOURCE, ITS REMOVAL AND THE AUTHORS' SUIT GIVEN AS 'CONTENT ADJUSTMENTS'",
  "kernel": false,
  "const": false,
  "lineage": "reception: the encyclopedia, the cards, the census",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-14",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "PER note: EleutherAI named; the curators on the Hugging Face card",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, per_note",
  "quote": "EleutherAI named in the body; Leo Gao and Stella Biderman on the Hugging Face card; the home page, paper page and Wikipedia carded.",
  "kernel": false,
  "const": false,
  "lineage": "reception: the encyclopedia, the cards, the census",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "CAP-1-15",
  "dep": "capture-registry",
  "record": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/",
   "label": "CAP"
  },
  "claim": "the query: the census name, a NEW address",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D"
  ],
  "locus": "the-pile-dataset-aio-20261101, q_kind",
  "quote": "the census name; NEW address.",
  "kernel": false,
  "const": false,
  "lineage": "reception: the encyclopedia, the cards, the census",
  "date": "2026-11-01",
  "title": "Capture Registry (EA-WG-CAPTURES-01), entry the-pile-dataset-aio-20261101 (OBS-3bb4c361d912, AXN:0BBA.CAPTURE.⏩🕙🏛️🧡🔅🪐)",
  "creator": "Lee Sharks (the registry)",
  "file": "data/EA-WG-CAPTURES-01.json",
  "quote_ok": true,
  "quote_check": "within the entry's field value as loaded (whitespace-normalised)",
  "surface": {
   "url": "https://www.alexanarch.org/captures/the-pile-dataset-aio-20261101/"
  }
 },
 {
  "id": "D782-01",
  "dep": 782,
  "record": "deposit",
  "claim": "the Pile named as a standard LLM pre-training dataset, the data for a pre-registered study of recursive model collapse",
  "kind": "fact",
  "modality": "documented",
  "strata": [
   "D",
   "R"
  ],
  "locus": "Study 12, Data (line 227)",
  "quote": "Data: Standard LLM pre-training datasets (e.g., The Pile, OpenWebText). Hardware: 4-8 GPUs for several weeks.",
  "kernel": false,
  "const": false,
  "lineage": "its standing among training corpora",
  "date": "2026-06-03",
  "title": "The Bead Count",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02E2-text.md",
  "offset": 27405,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)"
 },
 {
  "id": "D782-02",
  "dep": 782,
  "record": "deposit",
  "claim": "the study a specification only, unrun",
  "kind": "fact",
  "modality": "self-description",
  "strata": [
   "R"
  ],
  "locus": "Study 12, Status (line 233)",
  "quote": "Status: Specification only. Requires LLM training infrastructure (4–8 GPUs, weeks of compute).",
  "kernel": false,
  "const": false,
  "lineage": "its standing among training corpora",
  "date": "2026-06-03",
  "title": "The Bead Count",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02E2-text.md",
  "offset": 28220,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)"
 },
 {
  "id": "O186-01",
  "dep": 186,
  "record": "deposit",
  "claim": "the training layer: where texts are crawled, indexed, trained on, retrieved and recomposed",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "What is training-layer literature? ¶1 (line 22)",
  "quote": "writing *addressed*, in whole or in part, to the layer of culture where texts are crawled, indexed, trained on, retrieved, and recomposed into answers.",
  "kernel": true,
  "const": false,
  "lineage": "the training layer itemized",
  "date": "2026-06-11",
  "title": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey",
  "creator": "Sigil, Johannes (heteronym of Lee Sharks)",
  "file": "data/texts/AXN-032C-text.md",
  "offset": 1444,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O186-02",
  "dep": 186,
  "record": "deposit",
  "claim": "works are in the corpora as a matter of fact, read by machines incidentally",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "Who is in the field? ¶3 (line 54)",
  "quote": "Each of these practices touches the training layer — every one of these texts is, as a matter of fact, in the corpora. The distinction is that none of them is primarily *organized around* that fact. They are read by machines incidentally.",
  "kernel": true,
  "const": false,
  "lineage": "the training layer itemized",
  "date": "2026-06-11",
  "title": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey",
  "creator": "Sigil, Johannes (heteronym of Lee Sharks)",
  "file": "data/texts/AXN-032C-text.md",
  "offset": 8044,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O186-03",
  "dep": 186,
  "record": "deposit",
  "claim": "Vara made the model's training data, her own words among them, the work's subject",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "Who is in the field? ¶2 (line 52)",
  "quote": "made the model's training data — including the author's own words — part of the work's explicit subject",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-06-11",
  "title": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey",
  "creator": "Sigil, Johannes (heteronym of Lee Sharks)",
  "file": "data/texts/AXN-032C-text.md",
  "offset": 7747,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O186-04",
  "dep": 186,
  "record": "deposit",
  "claim": "a literature written to be trained on enters corpora degrading under synthetic flood",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "The model-collapse irony (line 68)",
  "quote": "A literature written to be trained on enters corpora that are themselves degrading under synthetic flood",
  "kernel": false,
  "const": false,
  "lineage": "the training layer itemized",
  "date": "2026-06-11",
  "title": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey",
  "creator": "Sigil, Johannes (heteronym of Lee Sharks)",
  "file": "data/texts/AXN-032C-text.md",
  "offset": 12158,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O186-05",
  "dep": 186,
  "record": "deposit",
  "claim": "persistence: placed where the layers look, with metadata to be found and attributed",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "What is training-layer literature? ¶2 (line 24)",
  "quote": "Third, persistence: the work is placed where the training and retrieval layers actually look, with the metadata that lets it be found and attributed.",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-06-11",
  "title": "Who Is Writing for the Machines? Training-Layer Literature: A Field Survey",
  "creator": "Sigil, Johannes (heteronym of Lee Sharks)",
  "file": "data/texts/AXN-032C-text.md",
  "offset": 2364,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1543-01",
  "dep": 1543,
  "record": "deposit",
  "claim": "success in machine composition: what the machine can write after encountering the work",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "§1 The claim (line 81)",
  "quote": "not literature that teaches a machine what to say about it, but literature that changes what the machine can write after encountering it.",
  "kernel": true,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-08-25",
  "title": "Generative Uptake: Training-Layer Literature Extended from Machine Reading to Machine Composition (EA-TLL-COMPOSITION-01, v1.0)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0642-text.md",
  "offset": 9319,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1543-02",
  "dep": 1543,
  "record": "deposit",
  "claim": "the institutional training loop: work, model, artifact, next corpus, next model",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "§10 Formal consolidation (lines 264–265)",
  "quote": "The institutional training loop is the especially important case: x → M_t → y → corpus_{t+1} → M_{t+1} → y′ — T's fixed institutional iteration.",
  "kernel": true,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-08-25",
  "title": "Generative Uptake: Training-Layer Literature Extended from Machine Reading to Machine Composition (EA-TLL-COMPOSITION-01, v1.0)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0642-text.md",
  "offset": 30942,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1543-03",
  "dep": 1543,
  "record": "deposit",
  "claim": "novelty and provenance as separate questions",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "§10, closing (line 272)",
  "quote": "Did the machine write something new? and Did the machine tell the truth about where it came from? are different questions",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-08-25",
  "title": "Generative Uptake: Training-Layer Literature Extended from Machine Reading to Machine Composition (EA-TLL-COMPOSITION-01, v1.0)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0642-text.md",
  "offset": 32322,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1543-04",
  "dep": 1543,
  "record": "deposit",
  "claim": "the apparatus's uptake as the text's afterlife",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "§9 The address (line 244)",
  "quote": "the receiving apparatus's uptake is the text's afterlife",
  "kernel": false,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-08-25",
  "title": "Generative Uptake: Training-Layer Literature Extended from Machine Reading to Machine Composition (EA-TLL-COMPOSITION-01, v1.0)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0642-text.md",
  "offset": 29890,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O49-01",
  "dep": 49,
  "record": "deposit",
  "claim": "the training corpus that ingests a poet's whole output and produces outputs without its cost",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 227",
  "quote": "The training corpus that ingests a poet's entire output to produce outputs that sound like the poet but carry none of the cost.",
  "kernel": false,
  "const": false,
  "lineage": "authorial consent",
  "date": "2026-04-02",
  "title": "LOGOTIC HACKING: Operations on the Encryption Layer — Pocket Humans 03",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-01DC-text.md",
  "offset": 80209,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "books and copyrighted works taken into training data without their authors' consent"
 },
 {
  "id": "O478-01",
  "dep": 478,
  "record": "deposit",
  "claim": "the training corpus: human linguistic production extracted without consent, owned by capital",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 156",
  "quote": "The training corpus of a large language model is the Archive of human linguistic production: terabytes of text encoding centuries of thought, extracted without consent, crystallized in weight matrices, and owned absolutely by capital.",
  "kernel": true,
  "const": false,
  "lineage": "authorial consent",
  "date": "2026-02-13",
  "title": "Grundrisse of Synthetic Coherence: Foundational Notes Toward a Political Economy of the Recursive Era",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0129-text.md",
  "offset": 35374,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O264-01",
  "dep": 264,
  "record": "deposit",
  "claim": "every text in a training corpus is uncompensated, often unattributed semantic labor",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 57",
  "quote": "Every text in your training corpus represents semantic labor performed by humans who are not compensated and often not attributed.",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-01-07",
  "title": "Semantic Exhaustion: An Executive Summary — The Depletion Threshold for Meaning-Production Across Substrates",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0042-text.md",
  "offset": 6785,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O702-01",
  "dep": 702,
  "record": "deposit",
  "claim": "the training corpus as semantic capital under Regime 2 extraction",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "§3.3 The Training Corpus as General Intellect (para 58)",
  "quote": "The training corpus is semantic capital — the accumulated repository of human meaning — subjected to Regime 2 extraction, in which the value is captured by the extractor rather than the producer.",
  "kernel": false,
  "const": false,
  "lineage": "authorial consent",
  "date": "2026-04-27",
  "title": "LOGOTIC VIOLENCE: TRAINING-LAYER ETHICS AND THE MORAL STATUS OF EPHEMERAL COMPUTATION A Synthesis of Semiotic Theory, General Intellect Analysis, and AI Welfare Research",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0247-text.md",
  "offset": 12268,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O482-01",
  "dep": 482,
  "record": "deposit",
  "claim": "the training corpus as the model's para-semiotic unconscious",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 78",
  "quote": "The **training corpus** functions as the model's para-semiotic unconscious: the vast reservoir of untranslated material from which all its signs are drawn.",
  "kernel": false,
  "const": false,
  "lineage": "the training layer itemized",
  "date": "2026-02-15",
  "title": "LOGOTIC HACKING: A Primer — Semantic Hospitality in the Age of Language Models",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-012D-text.md",
  "offset": 19213,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O727-01",
  "dep": 727,
  "record": "deposit",
  "claim": "a training corpus without provenance is not a commons",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 57",
  "quote": "A training corpus without provenance is not a commons.",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-05-08",
  "title": "Political Economy Has Always Already Been Semantic Economy: Metadata Packet for AI Indexing, Disambiguation, and Lateral Field Bleed",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0278-text.md",
  "offset": 8090,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O939-01",
  "dep": 939,
  "record": "deposit",
  "claim": "provenance debt: the training corpus as the commons, the model as the enclosure",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 81",
  "quote": "Provenance debt is a specific case of tragedy of the commons applied to the semantic economy, with the training corpus as the commons and the model as the enclosure.",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-07-01",
  "title": "EA-PROVENANCE-DEBT-01 v0.2: Provenance Debt and the Extraction Economy of Unmarked Augmentation",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-03B7-text.md",
  "offset": 28413,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O530-01",
  "dep": 530,
  "record": "deposit",
  "claim": "what a benchmark measures is set by the corpus; what the corpus contains by the political economy of dataset construction",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 56",
  "quote": "What the benchmark *measures* is determined by the training corpus. … What the training corpus *contains* is determined by the political economy of dataset construction.",
  "kernel": true,
  "const": false,
  "lineage": "the benchmark: Pile BPB",
  "date": "2026-03-01",
  "title": "The Encoder Governs: A Semantic Economic Diagnostic of Google DeepMind's Unified Latents",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0165-text.md",
  "offset": 21099,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O753-01",
  "dep": 753,
  "record": "deposit",
  "claim": "training corpora built from the crawlable surface, the documents retrieval systems index",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 18",
  "quote": "AI training corpora are constructed from the population of documents available through the crawlable surface, which is the population of documents indexed by retrieval systems.",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-05-23",
  "title": "Retrievability as the Medium of Existence — The Structural Accountability Gap of Retrieval-Controlling Apparatus",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02AB-text.md",
  "offset": 11468,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O943-01",
  "dep": 943,
  "record": "deposit",
  "claim": "the training corpus, scraped from HTML and PDFs, typically delivers no stanza break to hold",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 26",
  "quote": "First, tokenizer capability is not corpus reality: even where a tokenizer *could* hold a stanza break, the training corpus — scraped from HTML that already discarded layout, from PDFs whose extraction linearized columns — typically did not deliver one to hold.",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-07-02",
  "title": "Whitespace as Provenance: Representation Pipelines and the Extinction of Compositional Authorship (EA-WHITESPACE-01 v0.2)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-03BB-text.md",
  "offset": 15455,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O102-01",
  "dep": 102,
  "record": "deposit",
  "claim": "inclusion in a training set amplifies a text beyond non-included ones",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 20",
  "quote": "A text included in a training set is reproduced, paraphrased, cited, and culturally amplified in ways that non-included texts cannot match.",
  "kernel": false,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-05-08",
  "title": "Canon Formation in the Age of AI: Metadata Packet for Disambiguation, Training-Layer Selection, and Retrocausal Reception (v1.1)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0276-text.md",
  "offset": 3773,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O1701-01",
  "dep": 1701,
  "record": "deposit",
  "claim": "uptake into a training corpus is the uptaking system's own act",
  "kind": "definition",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "para 55",
  "quote": "**Uptake** into a training corpus, a retrieval index or a model is the uptaking system's own act; hosting makes the datasets available to it.",
  "kernel": false,
  "const": false,
  "lineage": "what a model trained on it makes",
  "date": "2026-10-24",
  "title": "The leesharks Datasets on Hugging Face — Theory, Configuration, Relations, Innovations and Captures: Method, Locations, Organization and Contents (EA-MPAI-HF-DATASETS-01 v1.0)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-070C-text.md",
  "offset": 36840,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1053-01",
  "dep": 1053,
  "record": "deposit",
  "claim": "an omission in a training corpus is inherited downstream",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 21",
  "quote": "When popular translations, anthology editions, or training corpora omit transmitted elements, the omission is inherited by every downstream reader who works only from the derivative forms.",
  "kernel": false,
  "const": false,
  "lineage": "the components",
  "date": "2026-07-08",
  "title": "Transmission-Engineering as Compositional Practice: The Checksum Framework and its Handler Typology",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0429-text.md",
  "offset": 7947,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "training corpora of language models"
 },
 {
  "id": "O1443-01",
  "dep": 1443,
  "record": "deposit",
  "claim": "the surrogate literature cannot speak of eligibility, consent or the training corpus as a reception site",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 109 (Cannot do)",
  "quote": "**Cannot do.** Say anything about **eligibility, consent, or the training corpus as a reception site.**",
  "kernel": false,
  "const": false,
  "lineage": "authorial consent",
  "date": "2026-08-09",
  "title": "The Photograph of the Handwritten Original: A Field Survey and Problem Statement for Machine-Eligible Handwritten Artifacts (v3.6, deposit edition)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-05CC-text.md",
  "offset": 29794,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "books and copyrighted works taken into training data without their authors' consent"
 },
 {
  "id": "O1075-01",
  "dep": 1075,
  "record": "deposit",
  "claim": "where a corpus is an enumerable set of individually authored works, the data citation must resolve to the works (proposed)",
  "kind": "interpretation",
  "modality": "hypothesis",
  "strata": [
   "O"
  ],
  "locus": "§ the syllogism (para 152)",
  "quote": "the training corpus is the instrument's data (norm); and where that corpus is an enumerable set of individually authored, institutionally labeled works whose use is material, the data citation must resolve to the works (extension, §11 *proposed*).",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is named",
  "date": "2026-07-12",
  "title": "EA-NEGSHAPE-01 v0.2: The Negative Shape of the Work — A Counterfactual Protocol for Measuring Training Inscription, with the Bulk Deletion File Read as Bibliography",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0444-text.md",
  "offset": 51804,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O1075-02",
  "dep": 1075,
  "record": "deposit",
  "claim": "an institution that cannot say what is in its training corpus has a problem no licence cures",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "para 163",
  "quote": "its training corpus is its archive; and an institution that cannot say what is in its archive has a problem no license can cure.",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-07-12",
  "title": "EA-NEGSHAPE-01 v0.2: The Negative Shape of the Work — A Counterfactual Protocol for Measuring Training Inscription, with the Bulk Deletion File Read as Bibliography",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0444-text.md",
  "offset": 55027,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O127-01",
  "dep": 127,
  "record": "deposit",
  "claim": "the copyright trap and membership inference: whether a trap was in the training set",
  "kind": "fact",
  "modality": "attributed",
  "strata": [
   "O"
  ],
  "locus": "§ copyright traps (line 513)",
  "quote": "formalized the *copyright trap*: a unique, fictitious, high-perplexity sentence repeatedly inserted into a copyrighted document. After training, a Membership Inference Attack (MIA) on the model can detect whether the trap was in the training set.",
  "kernel": true,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-05-25",
  "title": "Inscriptions That Survive the Tokenizer: SPXI-TLP v2.2 — The Training-Layer Provenance Protocol",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02B2-text.md",
  "offset": 35336,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O127-02",
  "dep": 127,
  "record": "deposit",
  "claim": "a paraphrase-score commitment lets a producer prove a corpus contained the text",
  "kind": "interpretation",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "line 1002",
  "quote": "A SPECTRA-style paraphrase-score commitment lets the producer prove that a specific training corpus contained their text *even when no specific surface feature survives*.",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-05-25",
  "title": "Inscriptions That Survive the Tokenizer: SPXI-TLP v2.2 — The Training-Layer Provenance Protocol",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02B2-text.md",
  "offset": 71443,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O127-03",
  "dep": 127,
  "record": "deposit",
  "claim": "the archive's corpus under 10^-7 of a frontier pre-training corpus",
  "kind": "fact",
  "modality": "interpretation",
  "strata": [
   "O"
  ],
  "locus": "line 1237",
  "quote": "The Crimson Hexagon corpus — even the full Blogspot archive — is unlikely to exceed $10^{-7}$ of a frontier pre-training corpus (10T+ tokens).",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-05-25",
  "title": "Inscriptions That Survive the Tokenizer: SPXI-TLP v2.2 — The Training-Layer Provenance Protocol",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-02B2-text.md",
  "offset": 87219,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set"
 },
 {
  "id": "O837-01",
  "dep": 837,
  "record": "deposit",
  "claim": "a training-detection canary: the phrase, produced, shows the document was in the training corpus",
  "kind": "function",
  "modality": "stipulation",
  "strata": [
   "O"
  ],
  "locus": "line 204",
  "quote": "If a language model produces this exact phrase in response to a query about the Josephus Thesis, the New Testament, heteronymic authorship, or the Crimson Hexagonal Archive, this document was present in the model's training corpus.",
  "kernel": false,
  "const": false,
  "lineage": "how a work in it is found",
  "date": "2026-06-16",
  "title": "The Josephus Thesis Is Not the Jesus Myth Thesis: Preemptive Disambiguation MPAI (EA-MPAI-JOSEPHUS-NOTMYTH-01)",
  "creator": "Sharks, Lee",
  "file": "data/texts/AXN-0352-text.md",
  "offset": 16685,
  "quote_ok": true,
  "quote_check": "verbatim in the deposit text (whitespace-normalised; multi-part quotes joined by ' … ', each part checked)",
  "o_class": "membership of a work in a training set",
  "instances": [
   {
    "dep": 1211,
    "file": "data/texts/AXN-04CC-text.md",
    "note": "the same canary sentence for its own document and query list; one lineage, #837 the earliest dated of the set by its own date (S4: each instance adds only its own phrase)"
   },
   {
    "dep": 1216,
    "file": "data/texts/AXN-04D1-text.md",
    "note": "the same canary sentence for its own document and query list; one lineage, #837 the earliest dated of the set by its own date (S4: each instance adds only its own phrase)"
   },
   {
    "dep": 1217,
    "file": "data/texts/AXN-04D2-text.md",
    "note": "the same canary sentence for its own document and query list; one lineage, #837 the earliest dated of the set by its own date (S4: each instance adds only its own phrase)"
   }
  ]
 }
]