LLMs capture substantial but selective and dimension-dependent components of human shared neural semantic representations, with larger models showing greater but incomplete overlap especially on agency, affect, and social dimensions.
theory of mind
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-bio.NC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Large language models selectively converge with human-shared neural semantic representations
LLMs capture substantial but selective and dimension-dependent components of human shared neural semantic representations, with larger models showing greater but incomplete overlap especially on agency, affect, and social dimensions.