In QAFactEval experiments, GPT-4 achieved a hallucination rate below 5%, while LLaMA 2 and DeepSeek exhibited hallucination rates between 20% and 25%.

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)

• hallucination rate concept