Semantic Drift Detector
Ki · Reasoning
The Problem
The term 'machine learning' now includes all AI methods, showing assumes definition stability over time.
The Operation
The model is directed to anchor key terms with operational definitions: what each includes, excludes, how to test. At each reasoning stage, re-extract working definitions and compare to anchors. Flag semantic drift: scope creep, narrowing, or meaning substitution. On silent drift, force re-anchoring: restate original, state divergence, choose one definition. Verify no conclusion rests on equivocation. Reason ambiguous terms is rejected.
The Structure
This ability runs on an anchor-drift-correct cycle that detects when reasoning has drifted from its reference point. The monitor runs continuously, checking for drift at each step.
If a key term is used later in the analysis with a subtly different meaning than when it was introduced, definitional consistency tracking was not performed.
Haki · Reasoning-Multi
Cross-Domain Suppression
In Haki mode, the API retrieves the primary ability first, then fans out to three synergy roles that compound its reasoning.
When retrieved in Haki mode, the primary ability is augmented with failure guards extracted from 3 abilities in different cognitive domains. Each guard blocks a specific reasoning failure the primary alone wouldn't catch. A self-check forces verification before output. The result is cross-domain coverage that no single ability can reach alone.
Appears in Use Cases