Hallucination

Appears in 1 paper

The phenomenon where a language model generates plausible-sounding but factually incorrect or entirely made-up information.

As used in Paper 12 — Language Models are Few-Shot Learners →

The phenomenon where a language model generates plausible-sounding but factually incorrect or entirely made-up information. The model learns patterns from text, not ground truth, so sometimes generates false facts confidently.