LEN-0290
Variance / Entropy Lens
Measures structural variability across repeated or comparable evaluations and identifies divergence beyond expected bounds.
Primary Pattern Matches
Asymmetric Structure
A structural condition where comparable regions receive unequal rule, constraint, or authority application without a declared differentiation rule.
Lens Application
The Variance / Entropy Lens is a strong inspection mechanism for Asymmetric Structure because it highlights where comparable regions show unexpected spread, drift, or inconsistency in applied rules. Elevated variance can focus review on places where authority, constraints, or enforcement may be distributed unevenly without a declared differentiation rule.
Inspect For
- Uneven outcomes across comparable regions
- Rule application variance beyond expected bounds
- Repeated exceptions clustered in one area
- Authority or constraint drift over time
Avoid
Treat variance as a signal for focused inspection rather than as sufficient evidence that Asymmetric Structure is present.
Density Spike
A structural condition where nodes, edges, dependencies, decisions, or effects concentrate sharply within a localized region beyond declared thresholds.
Lens Application
As a primary inspection mechanism, the Variance / Entropy Lens helps examine whether dependencies, decisions, effects, or connections cluster unevenly in one region compared with comparable areas or prior evaluations. It supports spotting abnormal concentration, rising divergence, or entropy changes that may indicate a density spike.
Inspect For
- Localized variance above expected bounds
- Sharp entropy drops or instability shifts
- Repeated concentration in one node group or region
- Divergence from comparable baseline evaluations
Avoid
Treat elevated variance or entropy change as a signal for closer inspection, not as confirmation that a density spike is present.
Divergent Outputs
A structural condition where parallel evaluations under comparable scope and shared authority produce non-equivalent outputs.
Lens Application
The Variance / Entropy Lens is a strong inspection mechanism for Divergent Outputs because it examines whether comparable evaluations are producing unusually inconsistent outputs. It supports inspection of output spread, decision-path instability, and whether divergence appears tied to scope, authority, criteria, or execution differences.
Inspect For
- Repeated evaluations producing materially different outputs
- Output variance under shared scope or authority
- Inconsistent weighting, criteria, or interpretation
- Divergence that exceeds expected tolerance
Avoid
Treat high variance as an inspection signal, not conclusive evidence of Divergent Outputs, until expected tolerance, bounds, and contextual justification are checked.
Escalation Growth
A structural condition where impact, authority, scope, or consequence increases across sequential states without a declared limiting mechanism.
Lens Application
The Variance / Entropy Lens is a strong inspection mechanism for Escalation Growth because it highlights whether sequential evaluations become increasingly variable, less bounded, or harder to predict. It helps surface growth patterns where escalation is not merely increasing, but spreading beyond expected structural limits.
Inspect For
- Increasing spread between comparable outcomes
- Repeated exceptions that expand scope or authority
- Rising unpredictability across sequential states
- Missing thresholds, caps, or reset mechanisms
Avoid
Treat variance signals as inspection prompts, not confirmation that Escalation Growth is present.
Non-Deterministic Execution
A structural condition where equivalent inputs and declared constraints produce divergent outputs across executions.
Lens Application
The Variance / Entropy Lens is a primary inspection mechanism for Non-Deterministic Execution because it focuses directly on divergence across repeated or comparable executions. It helps surface whether equivalent inputs and constraints produce inconsistent outputs, variable decision paths, or entropy that exceeds expected tolerances.
Inspect For
- Output differences across repeated equivalent runs
- Drift in intermediate states or execution paths
- Variability not explained by declared constraints
- Entropy spikes across comparable evaluations
Avoid
Treat variance signals as inspection signals requiring context, not as confirmation that Non-Deterministic Execution has been established.
Secondary Pattern Matches
Constraints Underspecified
A structural condition where declared constraints are insufficient to eliminate ambiguity or multiple admissible states.
Lens Application
The Variance / Entropy Lens applies indirectly to Constraints Underspecified by comparing repeated or similar evaluations for unusual spread, instability, or entropy. It does not identify underspecification directly, but can highlight places where insufficient constraints may be allowing multiple admissible outcomes.
Inspect For
- High variance across comparable evaluations
- Divergent outputs from similar inputs
- Entropy concentrated around constraint boundaries
- Repeated instability after minor changes
Avoid
Treat variance as a signal for further constraint inspection, not confirmation that Constraints Underspecified is present.
Convergence Failure
A structural condition where sequential or parallel states fail to resolve into an equivalent or coherent structure under shared authority and constraints.
Lens Application
The Variance / Entropy Lens applies indirectly to Convergence Failure by showing where repeated or parallel evaluations produce more variability than expected under shared authority and constraints. It can help reveal whether failed resolution is associated with unstable comparisons, inconsistent transformations, or entropy accumulating across attempts to align structure.
Inspect For
- Divergent outputs from comparable inputs
- Repeated evaluations producing unstable results
- Variance increasing across alignment attempts
- Parallel states drifting beyond expected bounds
Avoid
Treat high variance as a signal for closer inspection, not confirmation that Convergence Failure is present.
Propagation Amplification
A structural condition where an effect, declaration, authority, constraint, or state increases in scope or intensity as it propagates beyond declared bounds.
Lens Application
As a secondary lens, the Variance / Entropy Lens supports inspection of Propagation Amplification by comparing repeated or comparable propagation events for increasing divergence. It can highlight whether a declaration, constraint, authority, or state becomes less stable, more variable, or more intense as it travels across contexts.
Inspect For
- Increasing variability across downstream interpretations
- Scope or intensity changes between comparable propagation paths
- Divergence from expected limits or declared bounds
- Repeated outcomes becoming less predictable over distance
Avoid
Treat variance as a signal for further inspection rather than evidence that Propagation Amplification has occurred.
Unconstrained Expansion
A structural condition where a region, process, authority, or effect expands without governing constraints limiting growth.
Lens Application
As a secondary lens, the Variance / Entropy Lens supports inspection of Unconstrained Expansion by surfacing where expansion is producing increasing variation, drift, or disorder across repeated evaluations. Rising entropy may indicate that growth is outpacing coordination, constraint, or normalization, but it should be treated as a supporting inspection signal rather than confirmation.
Inspect For
- Increasing spread across comparable cases
- Divergent outcomes from similar inputs
- Escalating inconsistency over time
- Loss of stable reference bounds
Avoid
Treat variance as a supporting inspection signal, not as proof of Unconstrained Expansion without examining the governing constraints and surrounding structure.
Related Issues
Action Changed Something Else Too
An AI or agent action makes the requested change but also changes another object, field, file, state, rule, or workflow element that was not supposed to change.
Action Triggered by Confidence Score
A confidence score, certainty label, risk level, or probability-like value triggers an action without enough approval, calibration, or authority control.
Agent Keeps Expanding the Task
The agent repeatedly expands the task, plan, scope, or next-step list instead of completing the declared work.
Agent Never Settles on Final Answer
The agent keeps revising, rechecking, planning, or branching instead of converging on a final answer or completed result.
Agent Permission Expands Over Steps
An agent begins with limited permission but gains, assumes, or exercises broader authority as the workflow continues.
AI Forgets Earlier Constraints
A constraint, instruction, preference, or decision that should persist through the task stops affecting later output.
Answer Has Too Many Paths
The answer presents too many possible paths, interpretations, options, or next steps without enough structure to choose among them.
Context Leaks Between Tasks
Context, assumptions, constraints, examples, files, or decisions from one task affect another task where they should not apply.
Diagnostic Area Has No Coverage
A known diagnostic area, failure mode, requirement, or review dimension has no Issue, check, rubric item, or workflow coverage.
Downstream Steps Magnify Hallucinated Claim
A hallucinated or unsupported claim from an AI output is reused by later workflow steps until it becomes more influential than the evidence supports.
Duplicate Fields With Same Meaning
The AI returns multiple fields, labels, sections, or structured elements that carry the same meaning and create ambiguity about which one should be used.
Duplicate Output Sections
The AI repeats sections, headings, blocks, or output areas in a way that creates redundancy, confusion, or downstream handling problems.
Early Model Output Gets Overweighted Downstream
An early AI output receives too much authority in later workflow steps, decisions, reviews, or generated artifacts.
Evaluation Rubric Has Coverage Gap
An evaluation rubric, grading standard, or review checklist leaves part of the required evaluation space uncovered.
Format Rule Too Weak
The format instruction is too vague, incomplete, or optional to reliably produce output that satisfies the expected structure.
Human Review and Automation Disagree
A human review result and an automated AI or workflow result disagree without a declared rule for resolving the difference.
Invalid JSON Output
The AI returns malformed JSON or structured output that cannot be parsed.
Local Exception Grows Into Policy
A local exception, special case, or one-off allowance begins to function like a general policy.
Local Rule Spreads to Broader Cases
A rule intended for one local case, file, context, user, workflow, or exception begins affecting broader cases.
Merge Step Leaves Unresolved Differences
A merge, reconciliation, or consolidation step combines outputs or reviews but leaves important differences unresolved.
Missing Fallback for Unavailable Information
The task does not declare what the AI should do when required information, sources, tools, fields, or evidence are unavailable.
Model Output Triggers Unapproved Action
AI output causes, recommends, or triggers an action that has not passed the required approval, permission, or authority check.
Multiple Policies Say the Same Thing
Multiple policies, rules, or guidance documents express the same requirement, creating redundancy and uncertainty about which one governs.
One Prompt Carries Too Many Meanings
A single prompt carries too many meanings, goals, roles, constraints, or implied tasks for the AI to interpret consistently.
Output Breaks After Model Change
Output that previously worked begins failing after a model, mode, runtime, or product behavior changes.
Output Exceeds Length Limit
The AI output exceeds a declared length, token, word, character, section, field, or size limit.
Parallel Reviews Never Agree
Parallel AI, human, workflow, or tool reviews keep producing different results without resolving into a shared decision state.
Policy Exception Spreads Too Far
A narrow policy exception, allowance, or special case spreads beyond its intended scope and begins governing broader cases.
Prompt Behavior Changed Without Version Change
A prompt begins producing different behavior even though no prompt version, model version, workflow version, or declared dependency change is recorded.
Prompt Does Not Say What to Exclude
The prompt declares what to include but does not declare what should be excluded, allowing unwanted scope, sources, content, or actions into the result.
Prompt Has Too Many Valid Interpretations
The prompt allows too many reasonable interpretations, causing the AI to choose among valid paths without enough guidance.
Prompt Only Works After Retry
The prompt fails, misroutes, or produces an unusable response on one attempt but works after retry without a meaningful change to the input.
Repair Step Created New Breakage
A repair, correction, retry, or fix step addresses one problem but introduces a new failure elsewhere.
Repeated Constraints Create Confusion
Repeated constraints, instructions, limits, or exclusions make the task harder to interpret instead of clearer.
Results Vary Too Much
Repeated or comparable runs produce outputs that vary more than the task, workflow, or user can tolerate.
Retrieval Exceeds Evidence Limit
The AI retrieves, uses, cites, or considers more evidence than the task permits or more than the review surface can support.
Retry Makes the Problem Worse
A retry, repair attempt, regeneration, or follow-up instruction increases the error, expands the failure, or creates additional breakage instead of narrowing the problem.
Review Escalates Without Stop Condition
A review process keeps escalating, re-reviewing, or adding scrutiny without a declared condition for stopping.
Review Outcome Changes Unrelated Environment
A review result, approval, rejection, or classification changes state outside the environment, case, file, or workflow it was meant to govern.
Review Queue Becomes Bottleneck
A review queue, approval path, or validation stage accumulates too much work and begins blocking the workflow.
Review Rubric Missing Required Criteria
A review rubric, grading rule, evaluation checklist, or classification standard lacks criteria required to make the review reliable.
Risk Score Triggers Wrong Escalation
A risk score, severity label, confidence value, or threshold result triggers the wrong escalation path.
Risk Signal Escalates Beyond Evidence
A risk signal, warning, score, or concern escalates farther than the available evidence supports.
Routing Path Cycles Back to Start
A routing path sends the case back to the starting point or an earlier step without resolving the condition that caused the route.
Same Instructions Allow Different Outputs
The same instructions are broad or underspecified enough to allow materially different outputs that all appear compliant.
Same Rule Declared in Multiple Places
The same rule, constraint, instruction, or policy appears in multiple places, creating redundancy and possible drift.
Same Workflow Check Happens Twice
The same review, validation, approval, routing, or safety check occurs more than once in the workflow without a clear reason.
Severity Increases Without New Evidence
The severity, risk, confidence, or escalation level increases even though no new evidence has been added.
Similar Cases Route to Different Outcomes
Similar inputs, cases, prompts, or workflow states are routed to different outcomes without a declared difference that explains the split.
Single Field Carries Too Many Obligations
One field, label, score, status, or structured value is expected to carry too many meanings, decisions, or workflow obligations.
Single Step Carries Too Many Decisions
One prompt, workflow step, review stage, or agent action carries too many decisions for the system or user to evaluate cleanly.
Small Change Produces Large Downstream Effects
A small prompt, schema, policy, output, or workflow change creates unexpectedly large effects in downstream steps.
Small Error Spreads Into Large Failure
A small AI, output, routing, or workflow error propagates through later steps until it becomes a larger failure.
Small Issue Keeps Escalating
A small issue, warning, uncertainty, or correction keeps increasing in severity, scope, or workflow impact across later steps.
Task Has No Clear Limit
The task does not declare where the AI should stop, what is out of scope, or what counts as enough work.
Task Progress Is Lost Midway
The AI loses track of completed work, prior decisions, current position, or remaining steps before the task is finished.
Too Many Tool Calls
The AI or agent makes more tool calls, searches, retrievals, API calls, or integration actions than the task requires or permits.
Tool Exists but Required Inputs Are Missing
A usable tool or integration exists, but the AI or agent does not have the required inputs, permissions, fields, identifiers, or context needed to call it correctly.
Validation Result Changes on Retry
A validation, grading, review, classification, or pass/fail result changes after retry even though the input and declared validation rules did not change.
Workflow Loops Through Review Without Resolution
A workflow repeatedly sends work through review, repair, or escalation without reaching an approved, rejected, or otherwise resolved state.
Workflow Step Lacks Required Conditions
A workflow step can run, route, approve, reject, or continue without the required conditions being declared or checked.
Workflow Waits on Step That Waits Back
A workflow step waits for another step that also waits on the first step, creating a blocking loop.