Dontopedia

100k steps

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

100k steps has 18 facts recorded in Dontopedia across 4 references, with 4 live disagreements.

18 facts·12 predicates·4 sources·4 in dispute

Mostly:variant(3), rdf:type(2), condition(2)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (1)

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.

dependsOnDepends on(1)

Other facts (17)

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.

17 facts
PredicateValueRef
Variantgeneral-purpose model[4]
Varianttest the architecture's learning capacity on structured data[4]
Variantwant both[4]
Rdf:typeQuantitative Target[3]
Rdf:typeGoal[4]
Conditiongeneral-purpose model[4]
Conditiontest the architecture's learning capacity on structured data[4]
Outcomethis hurts[4]
Outcomedoesn't matter[4]
Instills Qa PatternCurrent Model[1]
Keep FromLarrikin Infection[2]
Fit foruseful and honourable work[2]
Forupright part in administration[2]
Recommended Actionmix TinyStories into the FineWeb data[4]
Recommended Ratio80/20[4]
Discouraged Actiondo a full epoch of pure domain shift[4]
Assessmentthis is a valid experiment[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.

18 undated
—
instillsQaPatternblah/watt-activation/part-163
ex:current-model
—
keepFromtrove-cooktown/prostitution
ex:larrikin-infection
—
fitFortrove-cooktown/prostitution
useful and honourable work
—
fortrove-cooktown/prostitution
upright part in administration
—
labelblah/training-and-evals/38
100k steps
—
typeblah/training-and-evals/38
ex:QuantitativeTarget
—
typeblah/watt-activation/144
ex:Goal
—
variantblah/watt-activation/144
general-purpose model
—
variantblah/watt-activation/144
test the architecture's learning capacity on structured data
—
variantblah/watt-activation/144
want both
—
conditionblah/watt-activation/144
general-purpose model
—
outcomeblah/watt-activation/144
this hurts
—
recommendedActionblah/watt-activation/144
mix TinyStories into the FineWeb data
—
recommendedRatioblah/watt-activation/144
80/20
—
discouragedActionblah/watt-activation/144
do a full epoch of pure domain shift
—
conditionblah/watt-activation/144
test the architecture's learning capacity on structured data
—
outcomeblah/watt-activation/144
doesn't matter
—
assessmentblah/watt-activation/144
this is a valid experiment

References (4)

4 references
  1. [1]Part 1631 fact
    ctx:discord/blah/watt-activation/part-163
  2. [2]Prostitution3 facts
    ctx:genes/trove-cooktown/prostitution
  3. [3]382 facts
    ctx:discord/blah/training-and-evals/38
    • full texttraining-and-evals-38
      text/plain3 KBdoc:agent/training-and-evals-38/63e5dd0a-e512-4eac-99ca-888173db5bcc
      Show excerpt
      [2026-03-09 12:03] foxhop.: (files: Screenshot_from_2026-03-09_08-01-18.png, Screenshot_from_2026-03-09_08-01-13.png) [2026-03-09 12:04] foxhop.: We are COOOOOOOOOOKING [2026-03-09 12:04] foxhop.: (files: Screenshot_from_2026-03-09_08-04-…
  4. [4]14412 facts
    ctx:discord/blah/watt-activation/144
    • full textwatt-activation-144
      text/plain3 KBdoc:agent/watt-activation-144/3be4aaf8-37ca-4d9d-bc3c-e22f23534527
      Show excerpt
      [2026-03-09 15:00] xenonfun: seems to really like it: step 100/64663 0.2% loss=5.3857 ppl= 218.3 lr=1.50e-05 668ms 12,265tok/s eta=719min step 200/64663 0.3% loss=4.7992 ppl= 121.4 lr=3.00e-05 667ms 12,277tok/s eta=…

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.

↑