LLaMA-2's hallucinations are primarily prompt-driven, characterized by high Prompt Sensitivity (PS) and low Model Variance (MV), meaning the model fails when prompts are suboptimal.

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)

• Prompt Sensitivity concept