AEO (Answer Engine Optimization): Optimization of content for inclusion in answer engine responses. A subset of LLMO’s Optimization layer.Calibration Error (delta): The expected absolute difference between stated confidence and true accuracy. High calibration error means the system is systematically overconfident or underconfident.Claim: A specific assertion about an entity. The atomic unit of truth in the LLMO system.Claim Record: A structured, machine-readable representation of a claim including its status, sources, freshness, and verification level.Conflict Resolution: The process of adjudicating between contradictory claims from multiple sources, governed by source hierarchy and recency.Cost per Verified Decision (CVD): Total loaded cost divided by verified decisions delivered. The pricing primitive for bounded-error cognition.Decision Envelope: The bounded context defining domain, risk tier, error bound, latency constraint, and applicable policies for a verified decision.EminenceFront: Systematic overconfidence bias that increases calibration error. Applies to both humans and models.Entity: A company, person, product, service, institution, domain, or concept that must be represented accurately across machine systems.Entity Record: The canonical machine-readable description of an entity including identity, claims, sources, and verification status.Evaluation Harness: The testing framework for measuring accuracy, calibration, refusal rates, freshness compliance, and provenance completeness.Freshness: Whether a claim is still current within its declared validity window.Freshness Window: The period during which a claim is considered current without re-verification.GEO (Generative Engine Optimization): Optimization of representation in generative systems. A subset of LLMO.Harness: The deterministic layer around stochastic models that provides routing, retries, policy enforcement, provenance, memory, evaluation, and auditability.Humans + Harness: The operating doctrine for safe, scalable machine-enabled execution. Models reason. Humans judge. The harness governs.LLMO (Large Language Model Optimization): The discipline of making entities, claims, and operational truth legible, trustworthy, and actionable inside machine reasoning systems.llmo.json: The reference implementation of the LLMO truth pack. A machine-readable artifact published at a domain root.Operations: The LLMO discipline governing internal machine-usable truth, workflow instrumentation, and governance.Optimization: The LLMO discipline governing external representation across LLM-driven systems.Orchestration: The LLMO discipline governing coordination of models, tools, humans, memory, and policies through deterministic workflows.Provenance: The trace from source to claim to representation. The chain of custody for a fact.Refusal State: A legitimate system output where the harness determines available truth is insufficient and declines to proceed.Replayability: The ability to re-execute a decision workflow from logged state and produce an auditable trace.Representation: How an entity is compressed and surfaced inside a model-mediated answer or workflow.SEO (Search Engine Optimization): Optimization of pages for search engine ranking. Predates and is subsumed by LLMO.Source: The origin of a claim, characterized by trust level, timestamp, ownership, and proximity to truth.Source Record: A structured, machine-readable description of a source including trust tier, type, and access history.Supersession: The replacement of an older claim by a newer authoritative claim from a source of equal or greater authority.Supersession Chain: The ordered sequence of claims that have replaced one another over time for a given entity attribute.Truth Pack: A portable, machine-readable artifact bundling entity records, claim records, source records, provenance, and verification levels. The atomic unit of trust in the LLMO protocol.Verified Decision Unit (VDU): A decision satisfying a defined domain, risk tier, error bound, and latency constraint. The atomic output of a Humans + Harness system.
