The study's methodology is limited by resource constraints that prevented testing of larger models, a focus on general-purpose tasks rather than domain-specific (e.g., medical, legal) tasks, a focus on short-to-medium-length responses, and a scope limited to open-source models up to 67B parameters.

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)

• Large Language Models concept