# What Do Visual-Head Discovery Scores Actually Find? Attention Source, Necessity, and Random-Control Pitfalls in Vision-Language Models

arXiv: https://arxiv.org/abs/2610.13111 · HTML: https://mlsanity.com/paper/2610.13111/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 39 of its 43 references that are in the map.
- First posted on arXiv 9 Oct 2026.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 30 of 39)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [Calibrate Before Use: Improving Few-Shot Performance of Language Models](https://mlsanity.com/paper/2102.09690/) (2021)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [Are We on the Right Way for Evaluating Large Vision-Language Models?](https://mlsanity.com/paper/2403.20330/) (2024)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs](https://mlsanity.com/paper/2406.16860/) (2024)
- [Are Sixteen Heads Really Better than One?](https://mlsanity.com/paper/1905.10650/) (2019)
- [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023)
- [Dissecting Recall of Factual Associations in Auto-Regressive Language Models](https://mlsanity.com/paper/2304.14767/) (2023)
- [Causal Abstractions of Neural Networks](https://mlsanity.com/paper/2106.02997/) (2021)
- [SpatialRGPT: Grounded Spatial Reasoning in Vision Language Models](https://mlsanity.com/paper/2406.01584/) (2024)
- [Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models](https://mlsanity.com/paper/2301.04213/) (2023)
- [Retrieval Head Mechanistically Explains Long-Context Factuality](https://mlsanity.com/paper/2404.15574/) (2024)
- [How to use and interpret activation patching](https://mlsanity.com/paper/2404.15255/) (2024)
- [Attribution Patching Outperforms Automated Circuit Discovery](https://mlsanity.com/paper/2310.10348/) (2023)
- [Teaching CLIP to Count to Ten](https://mlsanity.com/paper/2302.12066/) (2023)
- [Localizing Model Behavior with Path Patching](https://mlsanity.com/paper/2304.05969/) (2023)
- [See What You Are Told: Visual Attention Sink in Large Multimodal Models](https://mlsanity.com/paper/2503.03321/) (2025)
- [Towards Interpreting Visual Information Processing in Vision-Language Models](https://mlsanity.com/paper/2410.07149/) (2024)
- [AtP*: An efficient and scalable method for localizing LLM behaviour to components](https://mlsanity.com/paper/2403.00745/) (2024)
- [Learning to Predict Visual Attributes in the Wild](https://mlsanity.com/paper/2106.09707/) (2021)
- [Tell Your Model Where to Attend: Post-hoc Attention Steering for LLMs](https://mlsanity.com/paper/2311.02262/) (2023)
- [Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding](https://mlsanity.com/paper/2503.06287/) (2025)
- [Cross-modal Information Flow in Multimodal Large Language Models](https://mlsanity.com/paper/2411.18620/) (2024)
- [MIB: A Mechanistic Interpretability Benchmark](https://mlsanity.com/paper/2504.13151/) (2025)
- [Understanding Information Storage and Transfer in Multi-modal Large Language Models](https://mlsanity.com/paper/2406.04236/) (2024)

## Abstract (arXiv)

Attention heads that attend to image regions relevant to a question are increasingly used to steer, ground and interpret vision-language models (VLMs) and to reduce their hallucinations. They are mostly found by ranking heads with an attention- pattern score and validated by steering or masking them. We ask how the choices inside such a discovery score shape the heads it returns, and which of the two validations identifies them. On Qwen3-VL-8B-Instruct we vary the query type, the discovery corpus and the token positions whose attention is read (the attention source) in a full 2 x 3 x 2 design, and test all 12 head-sets for steerability (CERS: Contrastive Evidence-Region Steering score) and necessity (NCS: Necessity Con- trast Score) against layer-matched random head-sets. (1) The attention source and the query type reshape the selected heads far more than the corpus: changing the source keeps a median of 49 of the top-100 heads, the query 55 and the cor- pus 84, while two halves of the same data share 96-99. (2) Necessity follows the attention source: head-sets discovered from query-span attention are more nec- essary than last-token ones (difference 0.083, p=6 x 10^-10), and the two re- ferring + query-span sets lie 17-19 standard deviations above 250 layer-matched random sets (Holm-adjusted p=0.048). The referring + query-span effect re- plicates on several recent VLMs. (3) Steerability is not specific to the discovered heads: no discovered head-set steers better than its layer-matched random sets (rank p >= 0.14), and random sets drawn from the same or from any layer span the discovered sets' scores. Necessity therefore separates the discovery settings, while steering at this strength does not. We turn these findings into a protocol for reporting visual-head discoveries.