Dontopedia

Pairwise L1024

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

Pairwise L1024 has 20 facts recorded in Dontopedia across 3 references.

20 facts·18 predicates·3 sources

Mostly:has final ppl(2), has iterations per second(2), achieves best loss during training(1)

Maturity scale raw canonical shape-checked rule-derived certified

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.

inProgressAfterIn Progress After(1)

sharesProgressLogShares Progress Log(1)

Other facts (20)

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.

20 facts
PredicateValueRef
Has Final Ppl78.9[1]
Has Final Ppl78.9[2]
Has Iterations Per Second56[2]
Has Iterations Per Second56[3]
Achieves Best Loss During Training3.505[1]
Avg Iterations Per Second56[1]
Has Final Avg Loss4.3677[1]
Has Increasing Clusters Over TimeTraining Progress[1]
Has Kwargs{}[1]
Has Seq Len1024[1]
Shows Decreasing Avg LossTraining Progress[1]
Shows Decreasing PplTraining Progress[1]
Total Training Time356.4[1]
Has Best Loss3.505[1]
Has Best Ppl33.3[1]
Has R Value0.642[2]
Completed BeforeAnchor V3 L1024[2]
Rdf:typeMode[3]
Has Length1024[3]
Has Tokens Per Second Approximate58000[3]

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.

achievesBestLossDuringTrainingblah/watt-activation/part-59
3.505
avgIterationsPerSecondblah/watt-activation/part-59
56
hasFinalAvgLossblah/watt-activation/part-59
4.3677
hasFinalPPLblah/watt-activation/part-59
78.9
hasIncreasingClustersOverTimeblah/watt-activation/part-59
ex:training-progress
hasKwargsblah/watt-activation/part-59
{}
hasSeqLenblah/watt-activation/part-59
1024
showsDecreasingAvgLossblah/watt-activation/part-59
ex:training-progress
showsDecreasingPPLblah/watt-activation/part-59
ex:training-progress
totalTrainingTimeblah/watt-activation/part-59
356.4
hasBestLossblah/watt-activation/part-59
3.505
hasBestPPLblah/watt-activation/part-59
33.3
hasIterationsPerSecondblah/watt-activation/part-60
56
hasRValueblah/watt-activation/part-60
0.642
hasFinalPPLblah/watt-activation/part-60
78.9
completedBeforeblah/watt-activation/part-60
ex:anchor-v3-l1024
typeblah/watt-activation/62
ex:Mode
hasLengthblah/watt-activation/62
1024
hasIterationsPerSecondblah/watt-activation/62
56
hasTokensPerSecondApproximateblah/watt-activation/62
58000

References (3)

3 references
  1. [1]Part 5912 facts
    ctx:discord/blah/watt-activation/part-59
  2. [2]Part 604 facts
    ctx:discord/blah/watt-activation/part-60
  3. [3]624 facts
    ctx:discord/blah/watt-activation/62
    • full textwatt-activation-62
      text/plain2 KBdoc:agent/watt-activation-62/a54e16c9-84e0-4825-8097-9ebef4dc861d
      Show excerpt
      [2026-03-07 14:44] xenonfun: L512/L1024/L2048 data [2026-03-07 14:50] xenonfun: (files: phase_diagram_r_vs_ppl.png, training_trajectories_r_vs_ppl.png, scaling_r_and_ppl_vs_seqlen.png) [2026-03-07 14:51] xenonfun: ``` +### Per-configuratio

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.