Attention Head Analysis

Follow this topic Ask about this topic

The field investigates the roles, interactions, and stability of attention heads, including their contributions to induction, retrieval, reasoning, and multimodal tasks. Common approaches include head ablation and activation patching to test causal effects, alongside probes and spectral or geometric analyses to identify patterns and redundancies. These methods support both interpretability and practical goals such as pruning, bias analysis, and efficient inference across language, vision, and multimodal models. A central challenge is that attention patterns or probe scores do not by themselves establish a head’s causal role: effects can depend on the task, prompt, architecture, and controls used.

Latest papers

82 papers16 papers in the last four weeks, up 45% on the four weeks before. 0.1% of all new papers.

Self-Attention as Transport: Limits of Symmetric Spectral Diagnostics
May 6, 2026Dominik Dahlem, Diego Maniloff, Mac MisiuraTransformer Interpretability Attention Head Analysis
Red Hat AI
Gradient Flow Structure and Quantitative Dynamics of Multi-Head Self-Attention
May 5, 2026Ayan PendharkarSoftmax Attention Attention Head Analysis
Borrowed Geometry: Cross-Distribution Head-Importance Fingerprints of Frozen Pretrained Gemma 4 31B
May 1, 2026Abay BektursunCross-Modal Learning Attention Head Analysis
Independent research
Learning to Route Queries to Heads for Attention-based Re-ranking with Large Language Models
Apr 27, 2026Yuxing Tian, Fengran Mo, Zhiqi Huang +2Attention Head Analysis Learning to Rank
Université de Montréal · Capital One · McGill University & MILA
HeadRank: Decoding-Free Passage Reranking via Preference-Aligned Attention Heads
Apr 19, 2026Juyuan Wang, Chenxing Wang, Yuchen Fang +8Attention Head Analysis Self-Attention
Weixin Group, Tencent, China · 2South China University of Technology, Guangzhou, China
What Is the Minimum Architecture for Prolepsis? Early Irrevocable Commitment Across Tasks in Small Transformers
Apr 16, 2026Éric JacopinTransformer Interpretability Attention Head Analysis
Cosmic AI
On the Invariants of Softmax Attention
Apr 6, 2026Wonsuk LeeSoftmax Attention Attention Head Analysis
Seoul National University & SK Hynix America
Compressible Softmax-Attended Language under Incompressible Attention
Apr 6, 2026Wonsuk LeeSoftmax Attention Attention Head Analysis
Seoul National University & SK Hynix America
HeSS: Head Sensitivity Score for Sparsity Redistribution in VGGT
Mar 26, 2026Yongsung Kim, Wooseok Song, Jaihyun Lew +3Sparse Attention Efficient ViTs
IPAI, Seoul National University · ECE, Seoul National University · AIIS, ASRI, INMC, ISRC, Seoul National University
Attention in Space: Functional Roles of VLM Heads for Spatial Reasoning
Mar 21, 2026Xueqi Ma, Shuo Yang, Yanbei Jiang +6Visual Spatial Reasoning Attention Head Analysis
The University of Melbourne · Australian National University · Fudan University +1
Locating Demographic Bias at the Attention-Head Level in CLIP's Vision Encoder
Mar 12, 2026Alaa Yasser, Kittipat Phunjanna, Marcos Escudero Viñolo +2Attention Head Analysis VLM Interpretability
Universidad Autónoma de Madrid, Spain · Instituto Superior Técnico, Universidade de Lisboa, Portugal · University of Bordeaux, France
Surgical Repair of Collapsed Attention Heads in ALiBi Transformers
Mar 10, 2026Palmer SchallonAttention Head Analysis Positional Encoding
Independent researcher
Interpretable-by-Design Transformers via Architectural Stream Independence
Mar 8, 2026Clayton Kerce, Alexis FoxTransformer Interpretability Neural Network Interpretability
Georgia Tech Research Institute
Engineering Verifiable Modularity in Transformers via Per-Layer Supervision
Mar 8, 2026J. Clayton KerceTransformer Interpretability Mechanistic Interpretability
Georgia Tech Research Institute
Linear Predictability of Attention Heads in Large Language Models
Mar 4, 2026Khalid Shaikh, Asmit Kumar Singh, Rebecca Christopher Dsouza +1Attention Head Analysis LLM Inference Acceleration
Georgia Institute of Technology · Santa Clara University
Systematic Evaluation of Single-Cell Foundation Model Interpretability Reveals Attention Captures Co-Expression Rather Than Unique Regulatory Signal
Feb 19, 2026Ihor KendiukhovMechanistic Interpretability Neural Network Interpretability
Department of Computer Science, University of Tübingen, Tübingen, Germany
The Anxiety of Influence: Bloom Filters in Transformer Attention Heads
Feb 19, 2026Peter BaloghAttention Head Analysis Attention Head Specialization
Quantifying LLM Attention-Head Stability: Implications for Circuit Universality
Feb 17, 2026Karan Bali, Jack Stanley, Praneet Suresh +1Transformer Interpretability Attention Head Analysis
Mila - Quebec Artificial Intelligence Institute, Montreal, Canada · McGill University, Montreal, Canada
Compressed Sensing for Capability Localization in Large Language Models
Feb 11, 2026Anna Bair, Yixuan Even Xu, Mingjie Sun +1Attention Head Analysis LLM Interpretability
Carnegie Mellon University
Garbage Attention in Large Language Models: BOS Sink Heads and Sink-aware Pruning
Jan 11, 2026Jaewon Sok, Jewon Yeom, Seonghyeon Park +2LLM Pruning Attention Head Analysis
Department of Rural Systems Engineering, Seoul National University · Graduate School of Data Science, Seoul National University · Department of Aerospace Engineering, Seoul National University
Emergence of Minimal Circuits for Indirect Object Identification in Attention-Only Transformers
Oct 28, 2025Rabin AdhikariAttention Head Analysis Transformer
Saarland University 66123 Saarbrücken
Performance-Efficiency Tradeoffs in Transformers: An Approximation Theory Perspective
Oct 4, 2025Ruoxi Yu, Haotian Jiang, Jingpu Cheng +3Softmax Attention Attention Head Analysis
Center for Data Science, Peking University · Institute for Functional Intelligent Materials, National University of Singapore · Department of Mathematics, National University of Singapore +1

May 6, 2026J/K move · Enter open · S save

Self-Attention as Transport: Limits of Symmetric Spectral Diagnostics

Dominik Dahlem, Diego Maniloff, Mac Misiura

Red Hat AI

Transformer Interpretability Attention Head Analysis Self-Attention LLM Hallucination Detection Spectral Methods

Open paper AskPDF arXiv

When a language model processes a hallucinated response, its attention routing tends to fail in one of two shapes: over-concentrating on a narrow set of positions, or spreading so diffusely that relevance is diluted, and the shape of the failure carries diagnostic signal. We study these shapes as a diagnostic characterization, computed from attention matrices under forced scoring of benchmark-labeled responses rather than during live generation. A widely used family of spectral methods analyzes the symmetric component of the degree-normalized attention operator, which governs transport capacity; we prove that every transpose-invariant spectral diagnostic of this operator is structurally orientation-blind (it cannot distinguish an operator from its transpose, and therefore cannot detect information-flow direction), with a converse to the blindness theorem bounding any Lipschitz diagnostic's transpose sensitivity by the asymmetry coefficient

G

. Pairing this with a closed-form bipartite-Cheeger landscape for canonical causal architectures, we show that uniform causal attention satisfies an

n

-independent floor

φ\ge 1/5

, while window attention pierces the floor as

O(w/n)

; failure modes are shape-different, not just value-different. This floor is an idealized-architecture benchmark, not an empirical attractor: the fraction of real attention heads that pierce it is itself an architectural signature. The resulting two-axis diagnostic (

φ

for capacity,

G

for direction) yields a falsifiable polarity prediction: bottleneck- and diffuse-dominated benchmarks should exhibit opposite polarity. Under length-controlled evaluation, transport features retain interpretable signal (0.62-0.84 LC-AUROC) across the tested decoder-only, encoder-only, and encoder-decoder models, with polarity reversing as predicted between HaluEval and MedHallu.