ChatGPT demonstrated consistency in mapping probabilities to specified ranks for ArchiMate elements: 100% of elements on rank 1 were specified as 'Very High' (compared to 75% for experts), 95% on rank 2 (compared to 57% for experts), and 86% for 'Medium' (compared to 41% for experts).

Authors

• Person: Benedikt Reitemeyer, Hans-Georg Fill Organization: arXiv
  Applying Large Language Models in Knowledge Graph-based ...

Sources

• Applying Large Language Models in Knowledge Graph-based ... arxiv.org Benedikt Reitemeyer, Hans-Georg Fill · arXiv via serper

Referenced by nodes (1)

• ChatGPT concept