Exploring the Limitations of Large Language Models in Compositional Relation Reasoning
–arXiv.org Artificial Intelligence
In Natural Language Processing (NLP), composition relations are essential for several reasons. They enable We present a comprehensive evaluation of large sophisticated understanding and generation of language by language models(LLMs)' ability to reason about allowing models to discern and construct complex relationships composition relations through a benchmark encompassing between entities within a sentence or across texts. For 1,500 test cases in English, designed instance, understanding that "uncle" refers to a composition to cover six distinct types of composition relations: familial relation since it means "one's parent's brother". Positional, Comparative, Personal, Mathematical, So it is a composition relation of "brother of" and "parent Identity, and Other.
arXiv.org Artificial Intelligence
Mar-4-2024
- Country:
- North America
- Dominican Republic (0.04)
- United States > Minnesota
- Hennepin County > Minneapolis (0.04)
- Canada > Ontario
- Toronto (0.04)
- Waterloo Region > Waterloo (0.04)
- Europe
- Portugal > Lisbon
- Lisbon (0.04)
- Italy > Tuscany
- Florence (0.04)
- Denmark > Capital Region
- Copenhagen (0.04)
- Belgium > Brussels-Capital Region
- Brussels (0.04)
- Portugal > Lisbon
- Asia > Middle East
- North America
- Genre:
- Overview (0.68)
- Research Report (0.64)
- Industry:
- Education > Curriculum > Subject-Specific Education (0.46)
- Technology: