Dontopedia

PG-19

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

PG-19 is standard eval dataset for byte-level LMs.

19 facts·18 predicates·2 sources

Mostly:consists of project gutenberg books pre1919(1), features clean english prose(1), has no markup noise(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (3)

Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.

partOfPart of(3)

Other facts (18)

The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.

18 facts
PredicateValueRef
Consists of Project Gutenberg Books Pre1919True[1]
Features Clean English ProseTrue[1]
Has No Markup NoiseTrue[1]
Has Total Size~11 GB[1]
Is All AsciiTrue[1]
Is Seventeen Times Bigger ThanFineweb Edu Byte File[1]
Is Standard Eval Dataset forByte Level Lms[1]
Preferred Over Fineweb Edu for ScaleTrue[1]
Presupposes Clean Data Suitability for LmsTrue[1]
Rdf:typeDataset[2]
Descriptionstandard eval dataset for byte-level LMs[2]
SourceProject Gutenberg[2]
Time Periodpre-1919[2]
Content Typeclean English prose[2]
Has Markup Noisefalse[2]
EncodingASCII[2]
Total Size~11 GB[2]
Size Comparison17× bigger than our FineWeb-Edu byte file[2]

Timeline

Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.

consistsOfProjectGutenbergBooksPre1919blah/watt-activation/part-447
ex:true
featuresCleanEnglishProseblah/watt-activation/part-447
ex:true
hasNoMarkupNoiseblah/watt-activation/part-447
ex:true
hasTotalSizeblah/watt-activation/part-447
~11 GB
isAllAsciiblah/watt-activation/part-447
ex:true
isSeventeenTimesBiggerThanblah/watt-activation/part-447
ex:fineweb-edu-byte-file
isStandardEvalDatasetForblah/watt-activation/part-447
ex:byte-level-lms
preferredOverFinewebEduForScaleblah/watt-activation/part-447
ex:true
presupposesCleanDataSuitabilityForLmsblah/watt-activation/part-447
ex:true
typeblah/watt-activation/445
ex:Dataset
labelblah/watt-activation/445
PG-19
descriptionblah/watt-activation/445
standard eval dataset for byte-level LMs
sourceblah/watt-activation/445
ex:project-gutenberg
timePeriodblah/watt-activation/445
pre-1919
contentTypeblah/watt-activation/445
clean English prose
hasMarkupNoiseblah/watt-activation/445
false
encodingblah/watt-activation/445
ASCII
totalSizeblah/watt-activation/445
~11 GB
sizeComparisonblah/watt-activation/445
17× bigger than our FineWeb-Edu byte file

References (2)

2 references
  1. [1]Part 4479 facts
    ctx:discord/blah/watt-activation/part-447
  2. [2]44510 facts
    ctx:discord/blah/watt-activation/445
    • full textwatt-activation-445
      text/plain3 KBdoc:agent/watt-activation-445/9fd17e7c-a213-4d83-a408-8377df103970
      Show excerpt
      [2026-03-20 19:10] xenonfun: ``` ⏺ Already at 2.20 BPB at step 458 with bs=256. The reference numbers from CLAUDE.md: ┌──────────────────────────────┬───────┬───────┬─────────────────────────────────────┐ │ Config

See also

Keep researching

Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.