Signals 4 · free daily AI digest

Using OCR Heads to Verbalize Image Semantics

Paper recorded by Signals 4 on 2026-09-16 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-16 on arXiv · recorded by Signals 4 on 2026-09-17

Category: cs.CL · 自然语言处理 · first seen 2026-09-17

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 h

Read on arXiv →

#22 most recent of 186 cs.CL papers we have recorded · ↑ newer: ReFigBench: Benchmarking Scientific Figure Reconstruction as Editable · ↓ older: Beyond frequency measures: Can contextual embeddings capture meaning c
Cite this page: Using OCR Heads to Verbalize Image Semantics: the #22 most recent of 186 cs.CL papers we have recorded (as of 2026-09-16). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/using-ocr-heads-to-verbalize-image-semantics.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.CL papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions