Dontopedia

Custom AI System

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

Custom AI System has 21 facts recorded in Dontopedia across 1 reference.

21 facts·21 predicates·1 sources

Mostly:avoids reinforcement learning(1), avoids reward hacking(1), avoids reward signals(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (7)

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.

areAvoidedAre Avoided(1)

builtCustomAiSystemBuilt Custom AI System(1)

describesCustomAiSystemDescribes Custom AI System(1)

drivesContinuousImprovementDrives Continuous Improvement(1)

enablesCloudAccessEnables Cloud Access(1)

isAvoidedTrainingMethodIs Avoided Training Method(1)

isOptionalProvisionIs Optional Provision(1)

Other facts (21)

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.

21 facts
PredicateValueRef
Avoids Reinforcement Learningtrue[1]
Avoids Reward Hackingtrue[1]
Avoids Reward Signalstrue[1]
Based on Model FamilyMistral Family[1]
Embodies Principled Engineeringtrue[1]
Emphasizes Continuous ImprovementError Feedback[1]
Fine Tuned Via DatasetsCurated Supervised Semantic Dialogue Datasets[1]
Focuses on Safetytrue[1]
Has Provisions for Local Runningtrue[1]
Implements Error Feedback LoopFuture Learning[1]
Is Iterative in Engineeringtrue[1]
Manifests Iterative Approachtrue[1]
Manifests Principled Approachtrue[1]
Manifests Transparent Approachtrue[1]
Prioritizes Human Centered Designtrue[1]
Prioritizes Safetytrue[1]
Prioritizes Semantic AlignmentHuman Centered Alignment[1]
Promotes Transparencytrue[1]
Runs on InfrastructureCloud Infrastructure[1]
Uses Semantic Supervisiontrue[1]
Uses Strict No Reinforcement Learning RegimeNo Rlhf Regime[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.

avoidsReinforcementLearningblah/omega/part-844
true
avoidsRewardHackingblah/omega/part-844
true
avoidsRewardSignalsblah/omega/part-844
true
basedOnModelFamilyblah/omega/part-844
ex:mistral-family
embodiesPrincipledEngineeringblah/omega/part-844
true
emphasizesContinuousImprovementblah/omega/part-844
ex:error-feedback
fineTunedViaDatasetsblah/omega/part-844
ex:curated-supervised-semantic-dialogue-datasets
focusesOnSafetyblah/omega/part-844
true
hasProvisionsForLocalRunningblah/omega/part-844
true
implementsErrorFeedbackLoopblah/omega/part-844
ex:future-learning
isIterativeInEngineeringblah/omega/part-844
true
manifestsIterativeApproachblah/omega/part-844
true
manifestsPrincipledApproachblah/omega/part-844
true
manifestsTransparentApproachblah/omega/part-844
true
prioritizesHumanCenteredDesignblah/omega/part-844
true
prioritizesSafetyblah/omega/part-844
true
prioritizesSemanticAlignmentblah/omega/part-844
ex:human-centered-alignment
promotesTransparencyblah/omega/part-844
true
runsOnInfrastructureblah/omega/part-844
ex:cloud-infrastructure
usesSemanticSupervisionblah/omega/part-844
true
usesStrictNoReinforcementLearningRegimeblah/omega/part-844
ex:no-rlhf-regime

References (1)

1 references
  1. [1]Part 84421 facts
    ctx:discord/blah/omega/part-844

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.