Dontopedia

Alpha Cognitive

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

Alpha Cognitive has 23 facts recorded in Dontopedia across 1 reference.

23 facts·23 predicates·1 sources

Mostly:assumes data fits in disk(1), contrasts with symbio run(1), covers epochs(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (4)

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.

bootingUpTrainingRunBooting Up Training Run(1)

impliesImpressedByResultsImplies Impressed by Results(1)

involvesMultipleEpochsInvolves Multiple Epochs(1)

relevantToCurrentWorkRelevant to Current Work(1)

Other facts (23)

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.

23 facts
PredicateValueRef
Assumes Data Fits in DiskPd Ssd[1]
Contrasts With Symbio Runstandard activations[1]
Covers Epochs3-4[1]
Depends on Accumulation Steps4[1]
Expects960m Tokens Processedcalculated[1]
Has Effective Batch32 sequences (8 × 4 accum steps)[1]
Has Symbiooff[1]
Has Total Steps30K[1]
Has Warmup Steps1000[1]
Processes Tokens Per Step~32K[1]
Processes Total Tokens≈960M[1]
Risks Out of Memory on Ramif full load[1]
Room to Increase Batch Sizeyes[1]
Takes Training Time5-10 hours[1]
Uses DataHistoric Chat Plus Gutenberg[1]
Uses DiskPd Ssd[1]
Uses GpuNvidia L4[1]
Uses Learning Rate Schedule3e-4 lr with cosine decay to 1e-5[1]
Uses Machine TypeG2 Standard 4[1]
Uses OsUbuntu 22 04[1]
Uses Standard Composed Activationsonly[1]
Uses Training ConfigCognitive 350m Json[1]
Uses ZoneUS East1 B[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.

assumesDataFitsInDiskblah/training-and-evals/part-31
ex:pd-ssd
contrastsWithSymbioRunblah/training-and-evals/part-31
standard activations
coversEpochsblah/training-and-evals/part-31
3-4
dependsOnAccumulationStepsblah/training-and-evals/part-31
4
expects960mTokensProcessedblah/training-and-evals/part-31
calculated
hasEffectiveBatchblah/training-and-evals/part-31
32 sequences (8 × 4 accum steps)
hasSymbioblah/training-and-evals/part-31
off
hasTotalStepsblah/training-and-evals/part-31
30K
hasWarmupStepsblah/training-and-evals/part-31
1000
processesTokensPerStepblah/training-and-evals/part-31
~32K
processesTotalTokensblah/training-and-evals/part-31
≈960M
risksOutOfMemoryOnRamblah/training-and-evals/part-31
if full load
roomToIncreaseBatchSizeblah/training-and-evals/part-31
yes
takesTrainingTimeblah/training-and-evals/part-31
5-10 hours
usesDatablah/training-and-evals/part-31
ex:historic-chat-plus-gutenberg
usesDiskblah/training-and-evals/part-31
ex:pd-ssd
usesGpublah/training-and-evals/part-31
ex:nvidia-l4
usesLearningRateScheduleblah/training-and-evals/part-31
3e-4 lr with cosine decay to 1e-5
usesMachineTypeblah/training-and-evals/part-31
ex:g2-standard-4
usesOsblah/training-and-evals/part-31
ex:ubuntu-22-04
usesStandardComposedActivationsblah/training-and-evals/part-31
only
usesTrainingConfigblah/training-and-evals/part-31
ex:cognitive-350m-json
usesZoneblah/training-and-evals/part-31
ex:us-east1-b

References (1)

1 references
  1. [1]Part 3123 facts
    ctx:discord/blah/training-and-evals/part-31

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.