Automated Attention Pattern Discovery at Scale in Large Language Models

• URL: http://arxiv.org/abs/2604.03764v1
• Date: Sat, 04 Apr 2026 15:32:25 GMT
• Title: Automated Attention Pattern Discovery at Scale in Large Language Models
• Abstract summary: We study repeated behaviors in large language models by mining completion scenarios in Java code datasets.<n>We introduce the Attention Pattern - Masked Autoencoder (AP-MAE), a vision transformer-based model that efficiently reconstructs masked attention patterns.
• Score: 13.589024367220198
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models have found success by scaling up capabilities to work in general settings. The same can unfortunately not be said for interpretability methods. The current trend in mechanistic interpretability is to provide precise explanations of specific behaviors in controlled settings. These often do not generalize, or are too resource intensive for larger studies. In this work we propose to study repeated behaviors in large language models by mining completion scenarios in Java code datasets, through exploiting the structured nature of code. We collect the attention patterns generated in the attention heads to demonstrate that they are scalable signals for global interpretability of model components. We show that vision models offer a promising direction for analyzing attention patterns at scale. To demonstrate this, we introduce the Attention Pattern - Masked Autoencoder(AP-MAE), a vision transformer-based model that efficiently reconstructs masked attention patterns. Experiments on StarCoder2 show that AP-MAE (i) reconstructs masked attention patterns with high accuracy, (ii) generalizes across unseen models with minimal degradation, (iii) reveals recurring patterns across inferences, (iv) predicts whether a generation will be correct without access to ground truth, with accuracies ranging from 55% to 70% depending on the task, and (v) enables targeted interventions that increase accuracy by 13.6% when applied selectively, but cause collapse when applied excessively. These results establish attention patterns as a scalable signal for interpretability and demonstrate that AP-MAE provides a transferable foundation for both analysis and intervention in large language models. Beyond its standalone value, AP-MAE also serves as a selection procedure to guide fine-grained mechanistic approaches. We release code and models to support future work in large-scale interpretability.

Related papers

• Unsupervised Semantic Segmentation Facilitates Model Understanding [5.902216784679335]
  Self-supervised learning (SSL) has produced a diverse landscape of vision transformers (ViTs)<n>We propose a simple and easily interpretable visualization protocol to make model understanding straightforward and intuitive for a broad audience.<n>We obtain novel insights into distinct positional biases and scaling behaviors, including strong boundary artifacts in DINOv3-Large model tokens.
  arXiv  Detail & Related papers  (2026-05-28T09:52:52Z)
• Diagnosing Generalization Failures from Representational Geometry Markers [8.403001493770427]
  We study generalization failures inspired by medical biomarkers.<n>We design and test network markers to probe structure, function links, identify prognostic indicators, and validate predictions in real-world settings.<n>This work demonstrates that representational geometry can expose hidden vulnerabilities, offering more robust guidance for model selection and AI interpretability.
  arXiv  Detail & Related papers  (2026-03-02T13:59:19Z)
• FSDAM: Few-Shot Driving Attention Modeling via Vision-Language Coupling [5.609178055761294]
  We present FSDAM, a framework that achieves joint attention prediction and caption generation with 100 annotated examples.<n> FSDAM achieves competitive performance on attention prediction, generates coherent, and context-aware explanations.<n>This work shows that effective attention-conditioned generation is achievable with limited supervision, opening new possibilities for practical deployment of explainable driver attention systems.
  arXiv  Detail & Related papers  (2025-11-16T17:50:30Z)
• Every Step Counts: Decoding Trajectories as Authorship Fingerprints of dLLMs [63.82840470917859]
  We show that the decoding mechanism of dLLMs can be used as a powerful tool for model attribution.<n>We propose a novel information extraction scheme called the Directed Decoding Map (DDM), which captures structural relationships between decoding steps and better reveals model-specific behaviors.
  arXiv  Detail & Related papers  (2025-10-02T06:25:10Z)
• Attention Basin: Why Contextual Position Matters in Large Language Models [16.11590856103274]
  We show that models systematically assign higher attention to items at the beginning and end of a sequence, while neglecting those in the middle.<n>We introduce Attention-Driven Reranking (AttnRank), a framework that estimates a model's intrinsic positional attention preferences.<n>AttnRank is a model-agnostic, training-free, and plug-and-play method with minimal computational overhead.
  arXiv  Detail & Related papers  (2025-08-07T08:08:08Z)
• Predicting the Performance of Black-box LLMs through Self-Queries [60.87193950962585]
  Large language models (LLMs) are increasingly relied on in AI systems, predicting when they make mistakes is crucial.<n>In this paper, we extract features of LLMs in a black-box manner by using follow-up prompts and taking the probabilities of different responses as representations.<n>We demonstrate that training a linear model on these low-dimensional representations produces reliable predictors of model performance at the instance level.
  arXiv  Detail & Related papers  (2025-01-02T22:26:54Z)
• Unsupervised Model Diagnosis [49.36194740479798]
  This paper proposes Unsupervised Model Diagnosis (UMO) to produce semantic counterfactual explanations without any user guidance. Our approach identifies and visualizes changes in semantics, and then matches these changes to attributes from wide-ranging text sources.
  arXiv  Detail & Related papers  (2024-10-08T17:59:03Z)
• Data-efficient Large Vision Models through Sequential Autoregression [58.26179273091461]
  We develop an efficient, autoregression-based vision model on a limited dataset. We demonstrate how this model achieves proficiency in a spectrum of visual tasks spanning both high-level and low-level semantic understanding. Our empirical evaluations underscore the model's agility in adapting to various tasks, heralding a significant reduction in the parameter footprint.
  arXiv  Detail & Related papers  (2024-02-07T13:41:53Z)
• Copilot4D: Learning Unsupervised World Models for Autonomous Driving via Discrete Diffusion [36.321494200830244]
  Copilot4D is a novel world modeling approach that first tokenizes sensor observations with VQVAE, then predicts the future via discrete diffusion. Our results demonstrate that discrete diffusion on tokenized agent experience can unlock the power of GPT-like unsupervised learning for robotics.
  arXiv  Detail & Related papers  (2023-11-02T06:21:56Z)
• Large Language Models with Controllable Working Memory [64.71038763708161]
  Large language models (LLMs) have led to a series of breakthroughs in natural language processing (NLP) What further sets these models apart is the massive amounts of world knowledge they internalize during pretraining. How the model's world knowledge interacts with the factual information presented in the context remains under explored.
  arXiv  Detail & Related papers  (2022-11-09T18:58:29Z)
• SparseBERT: Rethinking the Importance Analysis in Self-attention [107.68072039537311]
  Transformer-based models are popular for natural language processing (NLP) tasks due to its powerful capacity. Attention map visualization of a pre-trained model is one direct method for understanding self-attention mechanism. We propose a Differentiable Attention Mask (DAM) algorithm, which can be also applied in guidance of SparseBERT design.
  arXiv  Detail & Related papers  (2021-02-25T14:13:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.