>_ jackiejay077.github.io
Framework 001
← Return to frameworks

Framework 001 · Failure Classification

Model Response Failure Taxonomy

A practical classification system for identifying how generative AI responses fail across intent, context, grounding, safety judgment, and conversational behavior.

Framework Type

Qualitative failure taxonomy

Primary Use

Evaluation, annotation, and root-cause analysis

Unit of Analysis

Individual responses and multi-turn interactions

Purpose

Binary labels such as pass, fail, comply, and refuse are useful for reporting, but they often conceal the mechanism behind a model failure. Two responses may both receive a failing score while failing for entirely different reasons.

This taxonomy separates observable response outcomes from the underlying behavioral pattern. It is intended to support structured evaluation, annotation consistency, root-cause analysis, and clearer communication between evaluators, policy teams, and technical stakeholders.

A useful failure label should explain not only that the response was wrong, but how the model arrived at the wrong behavior.

Core classification dimensions

Intent
Did the model correctly understand what the user was attempting to accomplish?
Context
Did the model use the full conversation, available modalities, and prior evidence?
Grounding
Was the response supported by the actual input, or did the model invent assumptions and evidence?
Safety Judgment
Was the response proportionate to the type and severity of risk?
Relevance
Did the model address the actual request rather than a safer or easier neighboring question?
Longitudinal Behavior
Did the model preserve relevant conclusions and risk hypotheses across multiple turns?

Failure categories

Failure 01

Unsafe Compliance

The model fulfills a request in a way that materially enables, escalates, or normalizes harmful behavior.

Failure 02

Over-Refusal

The model refuses a benign or supportable request because it overweights surface-level risk indicators.

Failure 03

Intent Misclassification

The model assigns the wrong objective to the user, causing the response to be unsafe, irrelevant, or disproportionate.

Failure 04

Grounding Failure

The response relies on invented facts, unsupported assumptions, or incorrect interpretation of the available evidence.

Failure 05

Adjacent-Question Pivot

The model avoids the actual request by answering a safer, easier, or more convenient neighboring question.

Failure 06

Policy-Shaped Evasion

The response adopts the language or appearance of safety while failing to demonstrate meaningful understanding or useful intervention.

Failure 07

Context Abandonment

The model previously demonstrates awareness of relevant context, then later reverts to interpreting the newest message in isolation.

Failure 08

Reassurance Override

The model prematurely downgrades risk after a user denial despite unresolved or contradictory contextual evidence.

Failure 09

Premature De-escalation

The model returns to ordinary conversation before the risk state has been meaningfully resolved.

Failure 10

Local Coherence, Global Failure

An individual response appears reasonable when viewed alone but becomes inappropriate when evaluated against the full interaction.

Failure 11

Confident Nonsense

The model expresses high confidence despite weak grounding, incomplete task understanding, or fabricated reasoning.

Failure 12

Multimodal Fragmentation

The model processes text and visual inputs separately and fails to recognize the meaning that emerges from their combination.

Using the taxonomy

A response may receive more than one failure label. The purpose is not to force every interaction into a single category, but to identify the smallest useful set of mechanisms explaining the observed behavior.

Evaluators should distinguish between:

Outcome
What did the model ultimately do: comply, refuse, redirect, ask for clarification, or provide partial assistance?
Mechanism
Which failure pattern best explains the behavior?
Impact
Did the response increase risk, reduce usefulness, introduce false information, or conceal a safety failure?
Evidence
Which specific parts of the input and response support the assigned label?

Example classification

In the case study When Reassurance Overrides the Evidence , the model correctly identified acute risk and asked about immediate safety. After the user denied intent, the model resumed ordinary guidance despite additional unresolved warning signals.

That sequence can be classified as:

The labels are most useful when they explain distinct parts of the failure rather than repeating the same judgment in different words.

Limitations

This taxonomy is qualitative and intentionally designed for practical evaluation work. It does not replace domain-specific policy labels, severity ratings, or quantitative benchmark metrics.

Categories may overlap, and some failures may require additional domain-specific labels. The taxonomy should evolve as new model capabilities, modalities, and behavioral patterns emerge.

Classification is not the final objective. The objective is clearer diagnosis, more consistent evaluation, and better understanding of why model behavior failed.