Research map / Large language models / Language, safety and interpretability
Fine-grained Analysis of Brain-LLM Alignment through Input Attribution
Michela Proietti, Roberto Capobianco, Mariya Toneva
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| replace Integrated Gradients with SmoothGrad for Llama3.2-1B attributions (ablation) | Llama3.2-1B BA/NWP attribution · Harry Potter fMRI positional profile | oscillatory positional attribution pattern | Improved |
| shorten the context window (ablation) | Llama3.2-1B attribution · Harry Potter fMRI positional profile | oscillatory positional attribution pattern | Improved |
| reduce concatenated TRs from 4 to 1 or 2 (ablation) | BA attribution pipeline · Harry Potter fMRI (IoU, spread, features, positions) | BA vs NWP attribution differences | No clear effect |
| use SmoothGrad instead of Integrated Gradients for Mamba-1.4B (ablation) | Mamba-1.4B attribution · Harry Potter fMRI positional profile | primacy peak visibility | Improved |
| mask top-attributed words versus random words (ablation) | NWP and BA · Harry Potter fMRI masking | cross-entropy increase / drop in Pearson r | Improved |
Closest papers by meaning
- Brain-tuning Improves Generalizability and Efficiency of Brain Alignment in Speech Models (NeurIPS 2025 · similarity 0.90)
- Scaling and context steer LLMs along the same computational path as the human brain (NeurIPS 2025 · similarity 0.90)
- fMRI predictors based on language models of increasing complexity recover brain left lateralization (NeurIPS 2024 · 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 0.89)
- Improving Semantic Understanding in Speech Language Models via Brain-tuning (ICLR 2025 · similarity 0.88)
- Divergences between Language Models and Human Brains (NeurIPS 2024 · 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.
