Research map / Vision-language and multimodal / Human and embodied modeling
Disentangling syntax and semantics in the brain with deep networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use contextual GPT-2 layer 9 representations instead of word embeddings | Brain-score prediction on Narratives fMRI · Narratives fMRI | brain score (Pearson R) | Helped |
| Factorize representations into lexical and compositional classes | GPT-2-to-brain mapping · Narratives fMRI | cortical distribution of brain-score effects | Helped |
| Extract syntax by averaging activations over syntactically matched sentences | GPT-2-to-brain mapping · Narratives fMRI | cortical distribution of syntactic effects | Helped |
| Treat syntax and semantics as sharing a distributed substrate rather than separate modules | Brain representation of language · Narratives fMRI | cortical distribution of syntax and semantics | Helped |
| Use other transformer architectures instead of GPT-2 | Brain-score prediction · Narratives fMRI | brain score (Pearson R) | Hurt |
| Use random word sequences, length-matched random sentences, or randomly permuted words instead of the original sentences | GPT-2 layer 9 brain-score analysis · Narratives fMRI | brain score (Pearson R) | Hurt |
| Use fewer than eight syntactically matched sentences to estimate syntax | GPT-2 syntactic embeddings · Wikipedia sentences | cosine similarity stability | Hurt |
Closest papers by meaning
- Far from the Shallow: Brain-Predictive Reasoning Embedding through Residual Disentanglement (NeurIPS 2025 · similarity 0.89)
- Hierarchical Frequency Tagging Probe (HFTP): A Unified Approach to Investigate Syntactic Structure Representations in Large Language Models and the Human Brain (NeurIPS 2025 · similarity 0.88)
- The Semantic Hub Hypothesis: Language Models Share Semantic Representations Across Languages and Modalities (ICLR 2025 · similarity 0.87)
- Improving Semantic Understanding in Speech Language Models via Brain-tuning (ICLR 2025 · similarity 0.87)
- fMRI predictors based on language models of increasing complexity recover brain left lateralization (NeurIPS 2024 · similarity 0.87)
- TopoLM: brain-like spatio-functional organization in a topographic language model (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.