Dontopedia
Explore

Gpt 2

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

Gpt 2 has 12 facts recorded in Dontopedia across 9 references, with 1 live disagreement.

12 facts·11 predicates·9 sources·1 in dispute

Mostly:is benchmark(2), has14x more parameters(1), has15x more training steps(1)

Maturity scale raw canonical shape-checked rule-derived certified

Is Benchmarkin disputeisBenchmark

Has14x More Parametershas14xMoreParameters

Has15x More Training Stepshas15xMoreTrainingSteps

Was Trained onwasTrainedOn

  • 9B tokens[9]all time · Part 96

Trained ontrainedOn

  • 256 TPU v3 cores[8]all time · Part 94

Used Many CoresusedManyCores

  • 256[8]all time · Part 94

Is Referenced ArchitectureisReferencedArchitecture

  • true[6]all time · Part 83

Defines Pre Norm StyledefinesPreNormStyle

  • true[2]all time · Part 43

Achieves LossachievesLoss

  • ~4-5 nats[1]all time · Part 119

Defines ScaledefinesScale

Rdf:typerdf:type

Inbound mentions (10)

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.

comparesCompares(1)

is3700xSmallerThanIs3700x Smaller Than(1)

isComparedToIs Compared to(1)

isWithin1-6xOfIs Within1 6x of(1)

presupposesGpt2ArchitecturePresupposes Gpt2 Architecture(1)

referencesGpt2References Gpt2(1)

referencesModelReferences Model(1)

scalesToGpt2Scales to Gpt2(1)

statedDesireForStated Desire for(1)

targetsModelTargets Model(1)

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.

12 undated
—
achievesLossblah/watt-activation/part-119
~4-5 nats
—
definesPreNormStyleblah/resources/part-43
true
—
definesScaleblah/watt-activation/part-119
ex:model-size
—
has14xMoreParametersblah/watt-activation/part-326
ex:our-model
—
has15xMoreTrainingStepsblah/watt-activation/part-326
ex:our-model
—
isBenchmarkblah/watt-activation/part-343
ex:byte-efficiency
—
isBenchmarkblah/watt-activation/part-85
null
—
isReferencedArchitectureblah/watt-activation/part-83
true
—
typeblah/watt-activation/694
ex:LanguageModel
—
trainedOnblah/watt-activation/part-94
256 TPU v3 cores
—
usedManyCoresblah/watt-activation/part-94
256
—
wasTrainedOnblah/watt-activation/part-96
9B tokens

References (9)

9 references
  1. [1]Part 1192 facts
    customctx:discord/blah/watt-activation/part-119
  2. [2]Part 431 fact
    customctx:discord/blah/resources/part-43
  3. [3]Part 3262 facts
    customctx:discord/blah/watt-activation/part-326
  4. [4]Part 3431 fact
    customctx:discord/blah/watt-activation/part-343
  5. [5]Part 851 fact
    customctx:discord/blah/watt-activation/part-85
  6. [6]Part 831 fact
    customctx:discord/blah/watt-activation/part-83
  7. [7]watt-activation-6941 fact
    customctx:discord/blah/watt-activation/694
    • full textwatt-activation-694
      text/plain3 KBdoc:agent/watt-activation-694/aff7c81b-38cd-47d8-b758-cefc82bfcd50
      Show excerpt
      [2026-04-29 04:39] xenonfun: (files: Screenshot_2026-04-29_at_12.39.16_AM.png) [2026-04-30 05:21] xenonfun: had it come up with a more DGX optimized for training, said I wanted GPT-2 in 12hrs. don't know that I'll get that but its doing ok…
  8. [8]Part 942 facts
    customctx:discord/blah/watt-activation/part-94
  9. [9]Part 961 fact
    customctx:discord/blah/watt-activation/part-96

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.

↑