High-confidence hallucinations, which appear fluent and plausible but are factually incorrect, are particularly dangerous and difficult to detect automatically in AI systems.

Authors

• Person: Not available Organization: Frontiers
  Survey and analysis of hallucinations in large language models

Sources

• Survey and analysis of hallucinations in large language models www.frontiersin.org Frontiers via serper

Referenced by nodes (1)

• artificial intelligence concept