The paper investigates how general vision‑language models (VLMs) develop specialized optical character recognition (OCR) capabilities. By applying a causal intervention protocol, the authors identify sparse, stable OCR‑head sets in several VLMs and show that these heads largely overlap with textual retrieval/copy heads found in general VLMs. The study concludes that full‑sequence OCR functions as a dense multimodal copy‑and‑paste mechanism, and that when a VLM is fine‑tuned for OCR, it largely preserves the same head identities while redistributing their functional and causal strengths.
By Yuanxiang Huangfu, Hanmeng Zhong, Linqing Chen, Jeffrey Tiong Jee Hui
The paper introduces Visual Retrieval Heads (VRHs), a small fraction of attention heads in vision‑language models that are causally responsible for grounding text descriptions to image regions. By recasting head‑scoring methods and evaluating across eleven VLMs and five benchmarks, the authors show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect. VRHs generalize across various visual reference tasks, preserve output format while corrupting localization, and transfer causally across models sharing an LLM backbone.
By Chanho Park, Daehyeon Choi, Jihyun Lee, Minhyuk Sung
The paper demonstrates that vision‑language models (VLMs) possess a small set of attention heads, called Visual Retrieval Heads (VRHs), that are causally responsible for linking text prompts to specific image regions. By adapting head‑scoring techniques from language models, the authors identify VRHs as the heads whose attention from output prediction tokens, summed over the ground‑truth referent region, most reliably indicates causal grounding. Experiments across eleven VLMs and five referring‑expression benchmarks show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect, and that VRHs generalize across diverse visual tasks and transfer across models sharing an LLM backbone.
arXiv:2609.18345v1 Announce Type: new
Abstract: Large vision-language models (LVLMs) are commonly used with only a single text prompt as the input, or plus an image. In this paper, we demonstrate tha...
By Xiaomeng Wang, Martha Larson, Zhengyu Zhao
arXiv:2603. 06054v2 Announce Type: replace-cross Abstract: The use of Vision-Language Models (VLMs) in automated driving applications is becoming increasingly common, with the aim of leveraging their reasoning and generalisation capabilities to handle long-tail scenarios.
By Nikos Theodoridis, Reenu Mohandas, Ganesh Sistu, Anthony Scanlan, Ciar\'an Eising, Tim Brophy
Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces. We ask whether CoT requires continued access to image tokens, or whether it mainly operates over visual information already made available earlier in the forward pass.
arXiv:2605.27750v2 Announce Type: replace-cross
Abstract: Recent work has shown that Vision-Language Models (VLMs) used for optical character recognition (OCR) can generate plausible but visually uns...
By Antonia Karamolegkou, Nicolas Angleraud, Beno\^it Sagot, Thibault Cl\'erice
arXiv:2607. 12815v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces.
By Hiroto Osaka, Shohei Taniguchi, Gouki Minegishi, Kai Yamashita, Masahiro Suzuki, Yutaka Matsuo
arXiv:2606. 14703v1 Announce Type: cross Abstract: How a vision-language model internally solves the task of describing an image is far from obvious.
By Rohit Gandikota, David Bau
arXiv:2605. 20448v2 Announce Type: replace-cross Abstract: Vision-language models reliably name objects in a scene, but do they represent the 3D layout those objects inhabit?
By Animesh Maheshwari, Divyansh Sahu, Nishit Verma
The study investigates whether multimodal large language models (MLLMs) report bistable images, like the duck‑rabbit, in a manner similar to humans. Using the LLaVA family, researchers examined two dimensions: modulability (the influence of visual cues and linguistic priors) and exclusivity (whether responses commit to a single interpretation). Results show that both visual and linguistic manipulations shift reports in human‑consistent ways while maintaining predominantly exclusive responses, driven by competing image‑token representations and distinct bottom‑up and top‑down pathways.
By Ryota Takatsuki, Tomoki Doi, Amane Watahiki, Anil K. Seth, Hitomi Yanaka
arXiv:2602. 07025v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) exhibit puzzling failures in multi-object visual tasks, such as hallucinating non-existent elements or failing to identify the most similar objects among distractions.
By Daniele Savietto, Declan Campbell, Andr\'e Panisson, Marco Nurisso, Giovanni Petri, Jonathan D. Cohen, Alan Perotti