Emergence and Localisation of Semantic Role Circuits in LLMs
By: Nura Aljaafari, Danilo S. Carvalho, André Freitas
Potential Business Impact:
Finds how computers understand words and sentences.
Despite displaying semantic competence, large language models' internal mechanisms that ground abstract semantic structure remain insufficiently characterised. We propose a method integrating role-cross minimal pairs, temporal emergence analysis, and cross-model comparison to study how LLMs implement semantic roles. Our analysis uncovers: (i) highly concentrated circuits (89-94% attribution within 28 nodes); (ii) gradual structural refinement rather than phase transitions, with larger models sometimes bypassing localised circuits; and (iii) moderate cross-scale conservation (24-59% component overlap) alongside high spectral similarity. These findings suggest that LLMs form compact, causally isolated mechanisms for abstract semantic structure, and these mechanisms exhibit partial transfer across scales and architectures.
Similar Papers
Language over Content: Tracing Cultural Understanding in Multilingual Large Language Models
Computation and Language
Shows how computers understand different cultures.
Are formal and functional linguistic mechanisms dissociated in language models?
Computation and Language
Helps computers better understand and use language.
Unraveling the cognitive patterns of Large Language Models through module communities
Artificial Intelligence
Shows how computer brains learn like animal brains.