company · Korean semiconductor equipment manufacturer; 'Korea's ASML'
Methodology — read this first. This page borrows the cross-domain signal-harvesting mechanic that Cambridge Analytica pointed at private citizens, and points it back at public figures via public records — with the output oriented at citizen oversight infrastructure, not persuasion. Every metric below is a sourced count or ratio drawn from primary public records. No metric on this page measures psychology, personality, motive, or intent. The platform documents conduct, not interior state. See methodology for evidence-tier definitions.
S1 // Signature axes
6 numeric structural metrics · each is a sourced count or ratio · derived in code, no LLM interpretation
Axis
Score
Underlying counts
Documentation Density
97/100
354 sourced facts · log-scaled (100 ≈ 600 facts)
Conduct Contradiction
0/100
0 broken / 0 resolved promises
Inferential Dependency
58/100
206 inferential / 354 total facts · lower = more primary-sourced
5 eligible entities ranked by cosine similarity over the 6-axis signature vector · uses behavior only, not connection-density · JSON export ↗
Cluster membership is a structural observation about documented behavior, not a claim of social affinity or coordination. Two entities share a cluster when their 6-axis behavioral signatures point in the same direction, regardless of whether they have any connection to each other.
backward-looking · count-anchored · no forecasts · methodology-locked prose template
Each observation below is a sourced count across the 5 cluster members — what has been documented to date, never a prediction. The platform does not infer what HPSP or anyone else in the cluster will do next.
Cluster members with at least one closed loop documented
4 of 5 (80%)
Cluster members with at least one contradiction on record
1 of 5 (20%)
Cluster members with at least one documented broken promise
0 of 5 (0%)
Cluster members whose tracked money-flow chains show net inflow
0 of 5 (0%)
Cluster members whose record is >25% inferential (thinner documentation)
5 of 5 (100%)
What the prose template enforces:
Only past-tense observations of what HAS occurred, never future-tense predictions of what WILL occur.
Count-anchored ("M of K") so the operator can verify by going to each linked entity's profile.
Cluster-level only — never attributed to any single member.
Code-generated from raw counts; no LLM-written prose interpretation.
S4 // Citizen oversight surface
where the inversion happens · same data, pointed at accountability infrastructure
For an entity of this type, the following public oversight tools apply. Each link goes to a primary-source government or registered-nonprofit portal — no third-party intermediaries. Where this entity appears in the underlying records, you can pull the documentation yourself.
Government Accountability Office audits of federal programs and major contractor performance.
// What this page does NOT do
methodology humility · required on every behavioral signature
Does NOT measure psychology, personality, or any "Big Five" / OCEAN trait.
Does NOT predict what HPSP will do. Every axis is backward-looking.
Does NOT cluster entities for the purpose of guilt by association. Clustering, when added in a future phase, will use behavioral-signature similarity only, never co-affiliation.
Does NOT target private individuals. The page is gated to entities flagged as public figures or registered legal entities.
Does NOT generate prose interpretation of the metrics. Every score is a count or ratio with the underlying numbers visible — no LLM-written "what this means" paragraph.
Behavioral Signature · derived from data the platform currently holds for HPSP. Scores recompute on each request. Corrections / disputes: [email protected].
