Sorry Robot, Happy Human: Vision-Language Models Read Only One of Two Legible Typographic Layers
Paper recorded by Signals 4 on 2026-09-25 in cs.CL. Abstract reproduced from arXiv; link to the original below.
Published 2026-09-25 on arXiv · recorded by Signals 4 on 2026-09-28
Category: cs.CL · 自然语言处理 · first seen 2026-09-28
Abstract
Vision-language models (VLMs), despite their success in optical character recognition (OCR) tasks, are vulnerable to typographic attacks and have a fragile structure for images with multiple text layers. In this study, the DecoyBench dataset was created using the Decoy Font method. The dataset consists of 300 images, each containing text with sharp contour lines superimposed on another text with s
Read on arXiv →
Cite this page: Sorry Robot, Happy Human: Vision-Language Models Read Only One of Two Legible Typographic Layers: the #8 most recent of 267 cs.CL papers we have recorded (as of 2026-09-25). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/sorry-robot-happy-human-vision-language-models-read-only-one-of-two-legible-typo.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable:
papers.json
Get 4 AI signals a day by email — free.
Get 4 AI signals a day by email — free