Positive Joint Attribution Score (JAS) values indicate joint amplification of hallucinations by prompts and models.

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 concept