Dontopedia

Baseline

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-09.)

Baseline has 21 facts recorded in Dontopedia across 6 references, with 1 live disagreement.

21 facts·17 predicates·6 sources·1 in dispute

Mostly:rdf:type(3), did not use reduce lr on plateau(1), generated text at25k(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (12)

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.

measuredValueForModelMeasured Value for Model(7)

comparedPerformanceOfCompared Performance of(1)

isSlowerThanBaselineIs Slower Than Baseline(1)

leadsOnLossMetricsLeads on Loss Metrics(1)

servesAsServes As(1)

structuralRoleStructural Role(1)

Other facts (19)

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.

19 facts
PredicateValueRef
Rdf:typeModel[3]
Rdf:typeModel Component[4]
Rdf:typeMachine Learning Baseline[6]
Did Not Use Reduce Lr on Plateaunull[1]
Generated Text At25kOf a few of the scholars who considered the great person. In general, natural philosopher[1]
Is Early in Training At25ktrue[1]
Has Poor Generation Quality At25ktrue[1]
Has Avg100 Loss At25k3.3905[1]
Has Best Loss At25k1.8893[1]
Has Eval Loss At25k3.2769[1]
Has Total Params3170000[2]
Has Output BehaviorCross Prompt Collapse[3]
Generates Output for Multiple Promptssame grey output[3]
Number of Prompts Producing Same Output5[3]
Has Within Prompt Pairwise Cosine0.41[3]
Has Spatial Diversity Ratingdecent[3]
Has Metric Value0.41[3]
Has Mean Field Distance Value0.06[3]
Has R Global Final Value0.65[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.

21 undated
—
didNotUseReduceLrOnPlateaublah/watt-activation/part-38
null
—
generatedTextAt25kblah/watt-activation/part-38
Of a few of the scholars who considered the great person. In general, natural philosopher
—
isEarlyInTrainingAt25kblah/watt-activation/part-38
true
—
hasPoorGenerationQualityAt25kblah/watt-activation/part-38
true
—
hasAvg100LossAt25kblah/watt-activation/part-38
3.3905
—
hasBestLossAt25kblah/watt-activation/part-38
1.8893
—
hasEvalLossAt25kblah/watt-activation/part-38
3.2769
—
hasTotalParamsblah/watt-activation/part-372
3170000
—
labelblah/watt-activation/271
Baseline
—
typeblah/watt-activation/271
ex:Model
—
hasOutputBehaviorblah/watt-activation/271
ex:cross-prompt-collapse
—
generatesOutputForMultiplePromptsblah/watt-activation/271
same grey output
—
numberOfPromptsProducingSameOutputblah/watt-activation/271
5
—
hasWithinPromptPairwiseCosineblah/watt-activation/271
0.41
—
hasSpatialDiversityRatingblah/watt-activation/271
decent
—
hasMetricValueblah/watt-activation/271
0.41
—
hasMeanFieldDistanceValueblah/watt-activation/271
0.06
—
hasRGlobalFinalValueblah/watt-activation/271
0.65
—
typeblah/watt-activation/297
ex:ModelComponent
—
labelblah/watt-activation/374
Baseline
—
typebeam/684b0c2c-1042-46ec-af7a-469a189d44aa
ex:MachineLearningBaseline

References (6)

6 references
  1. [1]Part 387 facts
    ctx:discord/blah/watt-activation/part-38
  2. [2]Part 3721 fact
    ctx:discord/blah/watt-activation/part-372
  3. [3]27110 facts
    ctx:discord/blah/watt-activation/271
    • full textwatt-activation-271
      text/plain3 KBdoc:agent/watt-activation-271/46f23862-31e7-4923-a679-800cf12ff385
      Show excerpt
      [2026-03-13 17:36] xenonfun: ⏺ Here's where we are: CC model result: we swapped one collapse for another. - Baseline: All 5 prompts produce the same grey output (cross-prompt collapse). Within-prompt pairwise cos = 0.41 (decent spatia…
  4. [4]2971 fact
    ctx:discord/blah/watt-activation/297
    • full textwatt-activation-297
      text/plain2 KBdoc:agent/watt-activation-297/ad91f718-f038-464f-a6d2-ba91d77fe4e3
      Show excerpt
      [2026-03-14 05:23] xenonfun: 600K context, the UI crazy scroll thing and memory leaks are annoyin claude is sucking up 6GB now. [2026-03-14 05:24] xenonfun: ``` ⏺ Launched in rjs:longgen. This trains with longer context (512) and more steps…
  5. [5]3741 fact
    ctx:discord/blah/watt-activation/374
    • full textwatt-activation-374
      text/plain2 KBdoc:agent/watt-activation-374/cb7e122e-3f7b-4f32-9d43-e995d3de2717
      Show excerpt
      [2026-03-18 19:24] xenonfun: ⏺ VQ results in. Head-to-head comparison: ``` ┌─────────────────┬──────────┬───────────┬───────┐ │ Metric │ Baseline │ AnchorKAN │ VQ │ ├─────────────────┼──────────┼───────────┼───────┤ │ S1…
  6. [6]684b0c2c 1042 46ec Af7a 469a189d44aa1 fact
    ctx:claims/beam/684b0c2c-1042-46ec-af7a-469a189d44aa
    • full textbeam-chunk
      text/plain1 KBdoc:beam/684b0c2c-1042-46ec-af7a-469a189d44aa
      Show excerpt
      SVMs can be effective, especially with the right kernel and parameter tuning. ### 4. **Decision Tree Classifier** Decision Trees are simple yet effective for certain types of data and can be used as a baseline. ### 5. **Naive Bayes Classi…

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.

↑