GLOSSARY > HALLUCINATION
Hallucination
A phenomenon where an AI model generates highly plausible but factually incorrect or nonsensical information.
Hallucinations occur when an LLM prioritizes probabilistic fluency over factual accuracy. In an enterprise environment, decisions cannot be based on hallucinations. Mitigating them requires strict grounding via RAG, robust Model Evaluation pipelines, and mathematical constraints on the model's confidence thresholds.
Explore More