A new concept-centric GeoQA benchmark probes LLMs on abstraction, compositionality, and grounding of direction, distance, and topology, finding good abstraction, weak grounding, and a Mistral-family deficit.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Geo-Spatial Concept Probing of Large Language Models: Abstraction, Compositionality, and Grounding
A new concept-centric GeoQA benchmark probes LLMs on abstraction, compositionality, and grounding of direction, distance, and topology, finding good abstraction, weak grounding, and a Mistral-family deficit.