Dontopedia

Experiment

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-16.)

Experiment has 103 facts recorded in Dontopedia across 43 references, with 8 live disagreements.

103 facts·86 predicates·43 sources·8 in dispute

Mostly:rdf:type(6), has3 runs of(3), has step(3)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (43)

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.

isArtifactOfIs Artifact of(3)

usedInUsed in(3)

mentionsMentions(2)

stillRunningStill Running(2)

actionAction(1)

agreedToFitTruckWithWaterAsAgreed to Fit Truck With Water As(1)

aimsForAgentOrchestrationAims for Agent Orchestration(1)

calledByCalled by(1)

commitsToOvernightRunCommits to Overnight Run(1)

comparedToCompared to(1)

componentComponent(1)

componentsComponents(1)

confirmedByConfirmed by(1)

correctedTermFromCorrected Term From(1)

correspondsToCorresponds to(1)

declaresNoRepeatDeclares No Repeat(1)

dependencyDependency(1)

describesActionDescribes Action(1)

framesAsFrames As(1)

heldConstantInHeld Constant in(1)

implicatesUsefulnessImplicates Usefulness(1)

indicateOngoingProcessIndicate Ongoing Process(1)

indicatesCoherenceIndicates Coherence(1)

involvesActionInvolves Action(1)

isAIs a(1)

isGenuinelyNewIs Genuinely New(1)

isInIs in(1)

isRunningIs Running(1)

managedManaged(1)

measuredInMeasured in(1)

notCriminalNot Criminal(1)

onePerOne Per(1)

performsPerforms(1)

rdf:typeRdf:type(1)

shouldShould(1)

supportsMoreCodesSupports More Codes(1)

trueInTrue in(1)

Other facts (99)

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.

99 facts
PredicateValueRef
Rdf:typeActivity[35]
Rdf:typeOptimization Activity[36]
Rdf:typeResearch Activity[37]
Rdf:typeExperiment[39]
Rdf:typeProcedure[40]
Rdf:typeActivity[42]
Has3 Runs of80K steps[27]
Has3 Runs of20K steps[27]
Has3 Runs of40K steps[27]
Has StepTrack Best Combination[40]
Has StepRefine Search[41]
Has StepValidate Results[41]
SubsidisesGroups of Farmers[31]
SubsidisesGroups of Farmers North Qld[31]
ContextArchitecture[33]
ContextIntent Precision Evaluation[39]
PurposeFind Optimal Value[36]
PurposeIntent Precision Measurement[39]
RequiresActual Evaluation Logic[39]
RequiresPlaceholder Functions Replacement[41]
Must Be inRegistry[1]
Uses Shared Git RepoGit Repo[2]
Killed Itselfdue to my ignorance[3]
Implies Ongoing ResearchCashu Idea[3]
For IdeaRichards Cashu Idea[3]
Self Terminates Due toFoxhop Ignorance[3]
Will IntegrateWeights and Biases[4]
Has Completed Configs3[5]
Is Bandwidth LimitedTrue[5]
Has Total Configs4[5]
Presupposes Cleaned Corpus ExistsTrue[5]
Estimated Total Time24 min[5]
Motivated byshits and giggles[6]
Embodied by ArchitectureArchitecture[7]
Branches FromFineweb Base[8]
Let FinishSmoketest[9]
Provides Important InsightRotational Opt Lohe Enforcement[10]
Has Three Stagestrue[11]
Uses Real Codetrue[11]
Uses50x More Datatrue[11]
Takes Duration30-40 minutes[11]
Expected to Take30-40 minutes[11]
Launched inRjs Bigcorpus[11]
Uses5k Budgetnull[12]
Running Compiled WithCompiledrotationaladamw[13]
Uses Mx Compile onTraining Step[13]
References Lohe Ff Nv3As Comparison[13]
Is Sweep Over ModesWire Sweep[14]
Benchmarks Modulationstrue[14]
Total Runtime Estimate14 min[14]
Running Strongtrue[14]
Nearing Completiontrue[14]
Aims atHigh Throughput[15]
Involves Lr Schedulingnull[15]
Precedes Manifold AblationManifold Ablation[16]
Is Ongoingtrue[17]
Uses Percent Curriculumtrue[17]
Involves Curriculum LearningStages[17]
Has Remaining Configs5[17]
Involves Lr SweepLr Values[18]
Cleanly Isolatesnull[19]
Tests Improvement ofLohe Inductive Bias[19]
Isolates Question ofArchitecture[19]
Holds ConstantFusion Mechanics[19]
Lacks PortedUnit Tests[20]
Temporary WithoutUnit Tests[20]
Aims for Evidence ExtractionMixing Angle Extraction[21]
ShowsClever Wholes[22]
Is E24 Prefetch1kTrue[23]
Is ProbablySingle Most Important Experiment[24]
Deemed Most Importantnull[24]
Probably Most Importantnull[24]
Is1k Stepsnull[25]
Uses Lr Scaled Rotationnull[25]
Uses Cosine Lrnull[25]
Involves Patch SetsDiffusion[26]
Presupposes Cc MechanismModel[26]
Has Runs9[27]
Total Time42 min[27]
Uses5 SeedsNull[28]
Small Vesseltrue[29]
Repeated During Daynull[29]
Recorded inLocal Straits Paper[29]
Considered Thoroughly SatisfactoryParty Consensus[30]
Would Not Cost Country Muchnull[32]
Ranked As"single most important"[33]
Temporal Context"right now"[33]
Rdf:typeActivity[34]
InvolvesOver Salting[34]
Participant InvolvesCompanion[35]
Measured25 Percent Boost[37]
Has GoalImprove Intent Precision[38]
MeasuresPrecision Metric[38]
Has Threshold88%[38]
Requires ReplacementPlaceholder Evaluation Function[39]
DependencyEvaluation Function[39]
Has SectionRunning the Experiment[41]
Conducted onLlm Configuration[42]
Scheduled forweekend[43]

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.

mustBeInblah/omega/part-1202
ex:registry
usesSharedGitRepoblah/resources/part-49
ex:git-repo
killedItselfblah/tpmjs/part-48
due to my ignorance
impliesOngoingResearchblah/tpmjs/part-48
ex:cashu-idea
forIdeablah/tpmjs/part-48
ex:richards-cashu-idea
selfTerminatesDueToblah/tpmjs/part-48
ex:foxhop-ignorance
willIntegrateblah/watt-activation/part-11
ex:weights-and-biases
hasCompletedConfigsblah/watt-activation/part-60
3
isBandwidthLimitedblah/watt-activation/part-60
ex:true
hasTotalConfigsblah/watt-activation/part-60
4
presupposesCleanedCorpusExistsblah/watt-activation/part-60
ex:true
estimatedTotalTimeblah/watt-activation/part-60
24 min
motivatedByblah/watt-activation/part-94
shits and giggles
embodiedByArchitectureblah/watt-activation/part-118
ex:architecture
branchesFromblah/watt-activation/part-144
ex:fineweb-base
letFinishblah/watt-activation/part-190
ex:smoketest
providesImportantInsightblah/watt-activation/part-293
ex:rotational-opt-lohe-enforcement
hasThreeStagesblah/watt-activation/part-296
true
usesRealCodeblah/watt-activation/part-296
true
uses50xMoreDatablah/watt-activation/part-296
true
takesDurationblah/watt-activation/part-296
30-40 minutes
expectedToTakeblah/watt-activation/part-296
30-40 minutes
launchedInblah/watt-activation/part-296
ex:rjs-bigcorpus
uses5kBudgetblah/watt-activation/part-307
null
runningCompiledWithblah/watt-activation/part-313
ex:compiledrotationaladamw
usesMxCompileOnblah/watt-activation/part-313
ex:training-step
referencesLoheFFNv3blah/watt-activation/part-313
ex:as-comparison
isSweepOverModesblah/watt-activation/part-322
ex:wire-sweep
benchmarksModulationsblah/watt-activation/part-322
true
totalRuntimeEstimateblah/watt-activation/part-322
14 min
runningStrongblah/watt-activation/part-322
true
nearingCompletionblah/watt-activation/part-322
true
aimsAtblah/watt-activation/part-342
ex:high-throughput
involvesLrSchedulingblah/watt-activation/part-342
null
precedesManifoldAblationblah/watt-activation/part-357
ex:manifold-ablation
isOngoingblah/watt-activation/part-369
true
usesPercentCurriculumblah/watt-activation/part-369
true
involvesCurriculumLearningblah/watt-activation/part-369
ex:stages
hasRemainingConfigsblah/watt-activation/part-369
5
involvesLRSweepblah/watt-activation/part-412
ex:lr-values
cleanlyIsolatesblah/watt-activation/part-437
null
testsImprovementOfblah/watt-activation/part-437
ex:lohe-inductive-bias
isolatesQuestionOfblah/watt-activation/part-437
ex:architecture
holdsConstantblah/watt-activation/part-437
ex:fusion-mechanics
lacksPortedblah/watt-activation/part-590
ex:unitTests
temporaryWithoutblah/watt-activation/part-590
ex:unitTests
aimsForEvidenceExtractionblah/watt-activation/part-602
ex:mixing-angle-extraction
showsblah/watt-activation/part-630
ex:clever-wholes
isE24Prefetch1kblah/watt-activation/part-706
ex:true
isProbablyblah/watt-activation/part-304
ex:single-most-important-experiment
deemedMostImportantblah/watt-activation/part-304
null
probablyMostImportantblah/watt-activation/part-304
null
is1kStepsblah/watt-activation/part-319
null
usesLrScaledRotationblah/watt-activation/part-319
null
usesCosineLrblah/watt-activation/part-319
null
involvesPatchSetsblah/watt-activation/part-275
ex:diffusion
presupposesCCMechanismblah/watt-activation/part-275
ex:model
has3RunsOfblah/watt-activation/part-325
80K steps
hasRunsblah/watt-activation/part-325
9
totalTimeblah/watt-activation/part-325
42 min
has3RunsOfblah/watt-activation/part-325
20K steps
has3RunsOfblah/watt-activation/part-325
40K steps
uses5Seedsblah/watt-activation/part-434
ex:null
smallVesseltrove-cooktown/coloured-persons
true
repeatedDuringDaytrove-cooktown/coloured-persons
null
recordedIntrove-cooktown/coloured-persons
ex:local-straits-paper
consideredThoroughlySatisfactorytrove-cooktown/bloomfield
ex:party-consensus
subsidisestrove-cooktown/north-shore-full
ex:groups-of-farmers
subsidisestrove-cooktown/north-shore-full
ex:groups-of-farmers-north-qld
wouldNotCostCountryMuchrosie-reynolds-massacre-connection/full-archive-reingestion-2026-05-06-batch-0122
null
rankedAsblah/watt-activation/302
"single most important"
contextblah/watt-activation/302
ex:architecture
temporalContextblah/watt-activation/302
"right now"
typeseven-sisters-of-sleep/74
ex:Activity
involvesseven-sisters-of-sleep/74
ex:over-salting
typeseven-sisters-of-sleep/250
ex:Activity
labelseven-sisters-of-sleep/250
experiments
participantInvolvesseven-sisters-of-sleep/250
ex:companion
typebeam/73db6035-02e5-47c3-8506-076dd04c43ef
ex:OptimizationActivity
purposebeam/73db6035-02e5-47c3-8506-076dd04c43ef
ex:find-optimal-value
typebeam/14d0c405-2f52-4261-ad38-13be7b76835d
ex:ResearchActivity
labelbeam/14d0c405-2f52-4261-ad38-13be7b76835d
Experiment
measuredbeam/14d0c405-2f52-4261-ad38-13be7b76835d
ex:25-percent-boost
labelbeam/5da37977-83e8-48be-bdd8-808083c26ac7
Experiment
hasGoalbeam/5da37977-83e8-48be-bdd8-808083c26ac7
ex:improve-intent-precision
measuresbeam/5da37977-83e8-48be-bdd8-808083c26ac7
ex:precision-metric
hasThresholdbeam/5da37977-83e8-48be-bdd8-808083c26ac7
88%
typebeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:Experiment
requiresbeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:actual_evaluation_logic
contextbeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:intent_precision_evaluation
purposebeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:intent_precision_measurement
requiresReplacementbeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:placeholder_evaluation_function
typebeam/da4b2af2-543c-45ec-bf7e-4898bd1b0c59
ex:Procedure
hasStepbeam/da4b2af2-543c-45ec-bf7e-4898bd1b0c59
ex:track_best_combination
dependencybeam/b303fb91-c589-4be6-ba31-3846ba31cc29
ex:evaluation_function
hasStepbeam/11402421-e0dd-4257-81f5-18735667d931
ex:refine-search
hasStepbeam/11402421-e0dd-4257-81f5-18735667d931
ex:validate-results
requiresbeam/11402421-e0dd-4257-81f5-18735667d931
ex:placeholder-functions-replacement
hasSectionbeam/11402421-e0dd-4257-81f5-18735667d931
ex:running-the-experiment
typebeam/ce00563e-e1f2-4d44-9f0b-129b7d9b122f
ex:Activity
labelbeam/ce00563e-e1f2-4d44-9f0b-129b7d9b122f
Experiment
conductedOnbeam/ce00563e-e1f2-4d44-9f0b-129b7d9b122f
ex:LLM-configuration
scheduledForlme/fcd8360b-d286-419e-9148-ad1ffb3988c5
weekend

References (43)

43 references
  1. [1]Part 12021 fact
    ctx:discord/blah/omega/part-1202
  2. [2]Part 491 fact
    ctx:discord/blah/resources/part-49
  3. [3]Part 484 facts
    ctx:discord/blah/tpmjs/part-48
  4. [4]Part 111 fact
    ctx:discord/blah/watt-activation/part-11
  5. [5]Part 605 facts
    ctx:discord/blah/watt-activation/part-60
  6. [6]Part 941 fact
    ctx:discord/blah/watt-activation/part-94
  7. [7]Part 1181 fact
    ctx:discord/blah/watt-activation/part-118
  8. [8]Part 1441 fact
    ctx:discord/blah/watt-activation/part-144
  9. [9]Part 1901 fact
    ctx:discord/blah/watt-activation/part-190
  10. [10]Part 2931 fact
    ctx:discord/blah/watt-activation/part-293
  11. [11]Part 2966 facts
    ctx:discord/blah/watt-activation/part-296
  12. [12]Part 3071 fact
    ctx:discord/blah/watt-activation/part-307
  13. [13]Part 3133 facts
    ctx:discord/blah/watt-activation/part-313
  14. [14]Part 3225 facts
    ctx:discord/blah/watt-activation/part-322
  15. [15]Part 3422 facts
    ctx:discord/blah/watt-activation/part-342
  16. [16]Part 3571 fact
    ctx:discord/blah/watt-activation/part-357
  17. [17]Part 3694 facts
    ctx:discord/blah/watt-activation/part-369
  18. [18]Part 4121 fact
    ctx:discord/blah/watt-activation/part-412
  19. [19]Part 4374 facts
    ctx:discord/blah/watt-activation/part-437
  20. [20]Part 5902 facts
    ctx:discord/blah/watt-activation/part-590
  21. [21]Part 6021 fact
    ctx:discord/blah/watt-activation/part-602
  22. [22]Part 6301 fact
    ctx:discord/blah/watt-activation/part-630
  23. [23]Part 7061 fact
    ctx:discord/blah/watt-activation/part-706
  24. [24]Part 3043 facts
    ctx:discord/blah/watt-activation/part-304
  25. [25]Part 3193 facts
    ctx:discord/blah/watt-activation/part-319
  26. [26]Part 2752 facts
    ctx:discord/blah/watt-activation/part-275
  27. [27]Part 3255 facts
    ctx:discord/blah/watt-activation/part-325
  28. [28]Part 4341 fact
    ctx:discord/blah/watt-activation/part-434
  29. [29]Coloured Persons3 facts
    ctx:genes/trove-cooktown/coloured-persons
  30. [30]Bloomfield1 fact
    ctx:genes/trove-cooktown/bloomfield
  31. [31]North Shore Full2 facts
    ctx:genes/trove-cooktown/north-shore-full
  32. ctx:genes/rosie-reynolds-massacre-connection/full-archive-reingestion-2026-05-06-batch-0122
  33. [33]3023 facts
    ctx:discord/blah/watt-activation/302
    • full textwatt-activation-302
      text/plain3 KBdoc:agent/watt-activation-302/799ace9b-0832-4dc6-aafb-d5303a75e0d0
      Show excerpt
      [2026-03-14 07:32] xenonfun: ``` ┌───────────────────┬────────┬─────────────┬───────────────┐ │ Decoder │ Params │ Speed (B/s) │ Attention │ ├───────────────────┼────────┼─────────────┼───────────────┤ │ Softmax d=384
  34. [34]742 facts
    ctx:books/seven-sisters-of-sleep/74
    • full texttmp7we8ix39_seven-sisters-of-sleep_74
      text/plain2 KBdoc:agent/tmp7we8ix39_seven-sisters-of-sleep_74/d62c253d-0f63-413e-a624-3dd67eeb660b
      Show excerpt
      ther, that an accidental fire did that for the firm which in the other case it is affirmed that an extra glass of grog accomplished. There is nothing surprising in this, and either narrative may be true; most inventions 90 of this kind, li
  35. [35]2503 facts
    ctx:books/seven-sisters-of-sleep/250
    • full texttmpxt5n1hoy_seven-sisters-of-sleep_250
      text/plain2 KBdoc:agent/tmpxt5n1hoy_seven-sisters-of-sleep_250/c2188c48-49e2-4720-9fe5-ab7f0ff62520
      Show excerpt
      mmunication with the spiritual world, or had received a divine commission by which to govern. Most of the marvels of ancient times were no greater than the little experiments which the schoolboy now performs for his amusement and that of hi
  36. ctx:claims/beam/73db6035-02e5-47c3-8506-076dd04c43ef
  37. ctx:claims/beam/14d0c405-2f52-4261-ad38-13be7b76835d
  38. ctx:claims/beam/5da37977-83e8-48be-bdd8-808083c26ac7
  39. ctx:claims/beam/b303fb91-c589-4be6-ba31-3846ba31cc29
  40. ctx:claims/beam/da4b2af2-543c-45ec-bf7e-4898bd1b0c59
    • full textbeam-chunk
      text/plain1 KBdoc:beam/da4b2af2-543c-45ec-bf7e-4898bd1b0c59
      Show excerpt
      For each combination of weights, evaluate the intent precision using your test queries. Replace the placeholder function with your actual evaluation logic. ### Step 4: Track the Best Combination Keep track of the best combination of weigh
  41. ctx:claims/beam/11402421-e0dd-4257-81f5-18735667d931
    • full textbeam-chunk
      text/plain1 KBdoc:beam/11402421-e0dd-4257-81f5-18735667d931
      Show excerpt
      2. **Refine the Search**: If the initial search does not yield significant improvements, consider narrowing down the range or using more sophisticated optimization techniques. 3. **Validate Results**: Validate the results on a separate vali
  42. ctx:claims/beam/ce00563e-e1f2-4d44-9f0b-129b7d9b122f
  43. ctx:claims/lme/fcd8360b-d286-419e-9148-ad1ffb3988c5
    • full textbeam-chunk
      text/plain16 KBdoc:beam/fcd8360b-d286-419e-9148-ad1ffb3988c5
      Show excerpt
      [Session date: 2023/05/22 (Mon) 09:06] User: I was thinking of experimenting with some new cocktails this weekend. Do you have any recommendations for summer drinks that incorporate Hendrick's gin? Assistant: Hendrick's gin is a fantastic c

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.