Using OCR Heads to Verbalize Image Semantics

2026-09-16

AI Summary

研究通过识别对OCR至关重要的注意力头，揭示视觉语言模型从像素到语义的映射机制，并展示这些头在处理非文字概念时的应用潜力。

Abstract

How do VLMs map from pixels to semantics? To understand this general question, we focus on a narrow one: studying how VLMs perform optical character recognition (OCR). Across four models, we identify attention heads causally necessary for OCR, and discover that these are in fact general-purpose heads that output interpretable semantic features across all image tokens. For example, pointing these heads at an image token containing the word "bike" causes Qwen3-VL-8B to output "bike," but pointing them at a bird wing causes the model to output the token "feathers." We collapse these heads' attention weights into a single verbalization lens transformation that reveals interpretable semantic features in hidden states across all layers. When combined with projection to vocabulary space, we can obtain interpretable labels starting from layer 0, showing that image representations are in fact aligned with language in early layers. We find that we can also use the inverse of this transformation to edit non-word concepts, e.g., replacing a tractor with a revolver in a naturalistic image, providing causal evidence that this subspace is useful for more than just OCR. Our results are an example of how the study of specific mechanisms can shed light on broader interpretability problems.

Problem

• Visual Language Models
• Optical Character Recognition
• Interpretability
• Semantic Features

Innovation

• Attention Heads
• Semantic Features
• Verbalization Lens
• Vocabulary Space Projection
• Concept Editing