Concepts
The language of auditable Human–AI systems.
These pages are reference entries, not marketing copy. They name the objects LLMography needs in order to reconstruct a decision.
Concept
Agent provenance
The attribution of delegated actions to particular agents and roles.
Concept
Decision-trace completeness
How much of the path from goal to decision can actually be reconstructed.
Concept
Graph engineering
The design of stateful, branching execution graphs for agentic work.
Concept
Human–AI decision loop
A recurrent cycle in which humans and AI systems jointly observe, act, verify, and revise.
Concept
Human–AI decision trajectory
The reconstructed path through which a Human–AI system arrives at a decision.
Concept
Human oversight
The capacity of a person to inspect, redirect, approve, or stop an AI loop.
Concept
Interaction provenance
The record of prompts, replies, tool calls, interventions, and verification events.
Concept
LLMography
The science of reconstructing, measuring, and auditing Human–AI decision trajectories.
Concept
Loop autonomy
The degree to which a loop continues without human intervention.
Concept
Loop engineering
The design of iterative observe–act–verify–revise cycles in agentic systems.
Concept
Loop provenance
How observation, revision, feedback, and stopping conditions evolve inside a loop.
Concept
Prompt provenance
The origin and transformation history of prompts that enter a model.
Concept
Recursive AI dependency
When later AI steps rely on earlier AI outputs without independent human grounding.
Concept
Reproducibility
Whether a trajectory, or a relevant slice of it, can be replayed under stated conditions.
Concept
Stopping-condition reliability
Whether a loop stops for a stated, inspectable reason rather than drift or exhaustion.
Concept
System provenance
How humans, agents, models, tools, memory, and policy relate across a system.
Concept
Verification strength
How independently a claim or action is checked before it is accepted.