Dontopedia
Explore

8k Vocab

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-05-02.)

8k Vocab has 15 facts recorded in Dontopedia across 8 references.

15 facts·15 predicates·8 sources

Mostly:produces worse output(1), has ln(1), basis for random loss(1)

Maturity scale raw canonical shape-checked rule-derived certified

Produces Worse OutputproducesWorseOutput

  • signifigantly worse output at same training[5]all time · Part 232

Has LnhasLn

  • 8.987[8]all time · Part 119

Basis for Random LossbasisForRandomLoss

  • ln(8000)[2]all time · Part 118

Has Random Loss ApproxhasRandomLossApprox

  • 8.99[2]all time · Part 118

Is Limiter forisLimiterFor

Encodes Text in TokensencodesTextInTokens

  • 199M[7]all time · Part 89

Achieved Best PplachievedBestPpl

  • 84[1]all time · Part 46

Causes Huge ImprovementcausesHugeImprovement

Dropped Best Ppl FromdroppedBestPplFrom

  • 213[1]all time · Part 46

Chosen for ComparisonchosenForComparison

Provides Reference PointprovidesReferencePoint

Picks Up PatternspicksUpPatterns

Inbound mentions (2)

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.

isLnOfVocabSizeIs Ln of Vocab Size(1)

isRandomBaselineLossIs Random Baseline Loss(1)

Other facts (3)

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.

3 facts
PredicateValueRef
Causes Worse Output Than100k100k Vocab[5]
Causes Fragmentationnull[3]
Causes MemorizationSurface Patterns[4]

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.

15 undated
—
achievedBestPplblah/watt-activation/part-46
84
—
basisForRandomLossblah/watt-activation/part-118
ln(8000)
—
causesFragmentationblah/watt-activation/part-88
null
—
causesHugeImprovementblah/watt-activation/part-46
ex:best-ppl
—
causesMemorizationblah/watt-activation/part-32
ex:surface-patterns
—
causesWorseOutputThan100kblah/watt-activation/part-232
ex:100k-vocab
—
chosenForComparisonblah/watt-activation/part-43
ex:previous-test-series
—
droppedBestPplFromblah/watt-activation/part-46
213
—
encodesTextInTokensblah/watt-activation/part-89
199M
—
hasLnblah/watt-activation/part-119
8.987
—
hasRandomLossApproxblah/watt-activation/part-118
8.99
—
isLimiterForblah/watt-activation/part-88
ex:anchor-kan-model
—
picksUpPatternsblah/watt-activation/part-32
ex:structured-text
—
producesWorseOutputblah/watt-activation/part-232
signifigantly worse output at same training
—
providesReferencePointblah/watt-activation/part-43
ex:previous-test-series

References (8)

8 references
  1. [1]Part 463 facts
    customctx:discord/blah/watt-activation/part-46
  2. [2]Part 1182 facts
    customctx:discord/blah/watt-activation/part-118
  3. [3]Part 882 facts
    customctx:discord/blah/watt-activation/part-88
  4. [4]Part 322 facts
    customctx:discord/blah/watt-activation/part-32
  5. [5]Part 2322 facts
    customctx:discord/blah/watt-activation/part-232
  6. [6]Part 432 facts
    customctx:discord/blah/watt-activation/part-43
  7. [7]Part 891 fact
    customctx:discord/blah/watt-activation/part-89
  8. [8]Part 1191 fact
    customctx:discord/blah/watt-activation/part-119

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.

↑