analiza aiqbench
This study uses fMRI recordings from listeners to test how much of an LLM's brain-predictive power comes from interpretable semantic properties. It finds that most of the alignment overlaps with those properties, while some brain-predictive signal is unique to the semantic representation. For AI readers, it is a useful interpretability result, though it covers one model (LLaMA2-7B), one dataset, and seven participants.
Decomposing Brain-LLM Alignment with Interpretable Semantic Representations
arXiv:2610.21508v1 Announce Type: new Abstract: Large language models (LLMs) predict neural responses during naturalistic language comprehension, but the linguistic information underlying this brain alignment remains difficult to interpret. We use SCRIBE, an interpretable contextual semantic representation spanning 65 neurocognitively grounded semantic properties, as a lens for decomposing the brain alignment of one LLM, LLaMA2-7B. Using variance partitioning between LLaMA2-7B and SCRIBE on naturalistic fMRI data from seven participants listening to narrative stories (Moth Radio Hour), we separate brain-predictive variance shared between the two representations from variance unique to each. We find that shared variance is the dominant component across the whole brain and language-related regions, while both representations also retain significant unique contributions. We then examine whether the concreteness of individual semantic properties relates to shared and SCRIBE-unique brain-predictive variance. Concrete properties show greater shared variance with LLaMA2-7B than abstract properties across the whole brain and most examined regions, whereas abstract properties do not show greater SCRIBE-unique variance. Finally, property ablations reveal that semantic contributions extend across multiple dimensions, with cumulative effects on both shared and SCRIBE-unique variance. Together, these results show that, for LLaMA2-7B on this dataset, a substantial portion of brain-LLM alignment is associated with interpretable semantic information, while revealing complementary brain-predictive semantic information beyond the LLaMA2-7B representation.
Abstract:Large language models (LLMs) predict neural responses during naturalistic language comprehension, but the linguistic information underlying this brain alignment remains difficult to interpret. We use SCRIBE, an interpretable contextual semantic representation spanning 65 neurocognitively grounded semantic properties, as a lens for decomposing the brain alignment of one LLM, LLaMA2-7B. Using variance partitioning between LLaMA2-7B and SCRIBE on naturalistic fMRI data from seven participants listening to narrative stories (Moth Radio Hour), we separate brain-predictive variance shared between the two representations from variance unique to each. We find that shared variance is the dominant component across the whole brain and language-related regions, while both representations also retain significant unique contributions. We then examine whether the concreteness of individual semantic properties relates to shared and SCRIBE-unique brain-predictive variance. Concrete properties show greater shared variance with LLaMA2-7B than abstract properties across the whole brain and most examined regions, whereas abstract properties do not show greater SCRIBE-unique variance. Finally, property ablations reveal that semantic contributions extend across multiple dimensions, with cumulative effects on both shared and SCRIBE-unique variance. Together, these results show that, for LLaMA2-7B on this dataset, a substantial portion of brain-LLM alignment is associated with interpretable semantic information, while revealing complementary brain-predictive semantic information beyond the LLaMA2-7B representation.
Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
arXiv:2610.21508 [cs.AI]
(or arXiv:2610.21508v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.21508 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 303 cuvinte
Subiecte conexe
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · 8 oct. 2026
Cross-lingual brain-language model alignment is robust but challenges hierarchical and computational accounts
arXiv · 11 sept. 2026
Margin-Regularized Structured Semantic Alignment for Brain-Language Correspondence
arXiv · 20 aug. 2026
BAT-CLIP: Trimodal Alignment of Brain, Audio and Text
arXiv · 28 sept. 2026