LLM internal representations of placenames show measurable spatial autocorrelation, and sparse autoencoders extract a small number of geospatially interpretable features.
Advancing spatial reasoning in large language models: an in-depth evalu- ation and enhancement using the StepGame benchmark
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Geospatial Mechanistic Interpretability of Large Language Models
LLM internal representations of placenames show measurable spatial autocorrelation, and sparse autoencoders extract a small number of geospatially interpretable features.