Hallucination

Intermediate

Model-generated content that is fluent but unsupported by evidence or incorrect; mitigated by grounding and verification.

Full Definition

Model-generated content that is fluent but unsupported by evidence or incorrect; mitigated by grounding and verification.

Keywords

ungrounded output

Domains

Model Failure Modes

Related Terms

Large Language Model related to

A high-capacity language model trained on massive corpora, exhibiting broad generalization and emergent behaviors.

 Overgeneralization related to

Applying learned patterns incorrectly.

 Exposure Bias related to

Differences between training and inference conditions.

 Catastrophic Forgetting related to

Loss of old knowledge when learning new tasks.

 Distribution Shift related to

Train/test environment mismatch.

Concept Map

See how Hallucination connects to other concepts.

Open Knowledge Graph