# Which Attention Heads are like the Human Head? Not the Ones that Compute

arXiv: https://arxiv.org/abs/2609.37991 · HTML: https://mlsanity.com/paper/2609.37991/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 18 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Builds on (most cited first, 8 of 8)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [Function Vectors in Large Language Models](https://mlsanity.com/paper/2310.15213/) (2023)
- [Neither hype nor gloom do DNNs justice](https://mlsanity.com/paper/2312.05355/) (2023)
- [The LLM Language Network: A Neuroscientific Approach for Identifying Causally Task-Relevant Units](https://mlsanity.com/paper/2411.02280/) (2024)
- [Causality neq Invariance: Function and Concept Vectors in LLMs](https://mlsanity.com/paper/2602.22424/) (2026)
- [Large Language Models Show Signs of Alignment with Human Neurocognition During Abstract Reasoning](https://mlsanity.com/paper/2508.10057/) (2025)
- [When Language Models Lose Their Mind: The Consequences of Brain Misalignment](https://mlsanity.com/paper/2603.23091/) (2026)

## Abstract (arXiv)

Brain-AI alignment is often interpreted as a sign that model and brain perform similar computations. Whether the aligned units are causally involved in model computation is rarely checked. On an abstract pattern-completion task (AAABAAA rightarrow B), we compare LLM attention-head representations with human EEG and test how ablating those heads affects task performance. Alignment and causation dissociate: brain-aligned heads contribute to performance, but their removal is substantially less disruptive than removal of heads selected via attribution patching. We compare two head sets that prior interpretability work defines without reference to the brain: concept vectors (CVs), which represent abstract patterns across formats, and function vectors (FVs), selected for their contribution to correct-answer prediction. Brain alignment shows little association with FV scores, while its association with CV scores varies across models. Among brain-aligned heads, we find recurring attention profiles: one emphasizes distinctive elements (novelty heads), the other repeating elements (repetition heads). The novelty family tracks salience and attends to the same elements that humans look at, yet its removal is less damaging than random ablation on average. Repetition heads contribute modestly to performance and are associated with abstract-pattern representation (CVs). Across 17 models spanning 3B-72B parameters, FV-ranked removal is substantially more disruptive than brain-ranked removal. Brain alignment thus captures how the model reads the stimulus, and only faintly captures how it represents the pattern and solves the task.