Evaluation and reliability

Abstention

stable definition
Machine-readable Download Markdown

Definition

Abstention is an explicit outcome in which a model or evaluator declines to answer, act, or judge because its evidence or authority is insufficient. The abstention should route the case somewhere safe, such as a human reviewer, another evidence source, or a bounded request for clarification.

A useful abstention policy defines the trigger, the message, the destination, and what happens if the destination is unavailable. Evaluate coverage and selective risk together: how often the system abstains, and how error changes on the cases it still handles.

Abstention can be misused. A system may decline difficult cases that matter most, hide adverse findings behind uncertainty, or abstain differently across groups. Those patterns need measurement alongside overall accuracy.

Distinguish it from nearby terms

Abstention is a governed response to insufficient evidence or authority. A safety refusal rejects a request because policy prohibits it. A failure means the system could not complete the operation, whether or not it recognized that fact.

Check your understanding

A medical classifier abstains on every rare condition and reports excellent accuracy on the rest. The metric improves, but the routing path and the burden placed on those patients determine whether the system is safer.