ModelRefs / Abstention — AI Glossary
Abstention — AI Glossary
A model declining to answer due to uncertainty or knowledge gaps rather than policy violations, to avoid hallucination. Distinct from refusal (policy-driven).
Overview
Abstention is a quality-preserving behavior: a well-calibrated model should say 'I don't know' when it lacks reliable information rather than confabulate. Distinct from refusal (policy-driven). Improving abstention rates requires training on uncertainty signals. Evaluated by abstention precision/recall on factual QA benchmarks.
Reference details
| Topic | prompting |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Abstention — AI Glossary.
Frequently asked questions
What is Abstention?
A model declining to answer due to uncertainty or knowledge gaps rather than policy violations, to avoid hallucination.
What concepts are related to Abstention?
Closely related concepts include refusal, calibration, hallucination.