CORTEXA
← Browse
arxivcs.CVcs.CL2026-07-22

Test-Time Training for Modality Order Consistency in Vision-Language Models

Aditi Gupta, Yossi Gandelsman

We find that vision-language models are sensitive to a specific semantically irrelevant change: the order in which the image and question are presented. Across three models and three benchmarks, image first prompting consistently outperforms question-first prompting, revealing a repeatable modality order failure. We use this gap to design an order-consistent test-time training method. Our method substantially closes the modality-order gap across all evaluated settings. Surprisingly, it also yields consistent improvements in the stronger image-first branch over the baseline, hence bootstrapping both orderings toward mutual consistency. Activation patching localizes the ordering failure to a narrow mid-network region where representations diverge sharply between prompt orders. We find that the test-time training method repairs this misalignment across layers. Together, our results identify modality-order sensitivity as a circuit-level failure in VLMs and demonstrate that simple, asymmetric test-time adaptation can effectively mitigate it and even improve performance over the baseline.

View free PDFSource page

Related papers

arxivcs.CLcs.CV2026-07-02

Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning

Liyan Tang, Fangcong Yin, Greg Durrett

Large vision-language models can reason over multimodal inputs by generating textual chains of thought (CoT). A key capability exhibited in CoT reasoning is self-reflection: revisiting earlier decisions and correcting previous errors. However, existing LVLMs often fail to properl…

View free PDFSource page
arxivcs.IRcs.AIcs.CLcs.CV2026-07-06

Do All Visual Tokens Matter Equally? Object-Evidence Preserving Token Merging for Vision-Language Retrieval

Suhyeong Park, Junha Jung, Jungwoo Park, Jaewoo Kang

Multi-vector vision-language retrieval preserves fine-grained visual evidence through maximum-similarity late interaction, but dense image-side tokens make storage and scoring expensive. Existing token compression methods reduce this cost, yet they can remove or collapse object-…

View free PDFSource page
arxivcs.CVcs.CL2026-07-06

Does It Fail to See or Fail to Know? Attributing Errors in Vision-Language Models

Khang Nhat Hoang Vo, Artem Vazhentsev, Artem Shelmanov, Timothy Baldwin, Yova Kementchedjhieva

Vision-language models (VLMs) perform well on visual question answering with high-quality images but struggle when questions require knowledge beyond what is clearly and directly visible. In such settings, uncertainty quantification should not only indicate whether the model is l…

View free PDFSource page
arxivcs.CVcs.CLcs.LG2026-07-03

Pathways of Visual Information Flow in Vision-Language Models

Israfel Salazar, Stella Frank, Dan Oneata, Desmond Elliott, Constanza Fierro

We study how visual information is routed in vision-language models (VLMs). Using causal patching on controlled synthetic and natural datasets, we find that models rely on two distinct pathways to solve visual tasks: A direct pathway, where visual information is retained in image…

View free PDFSource page
arxivcs.CVcs.CLcs.CR2026-07-17

One Modality to Forget Them All: Enhancing Cross-Modal Unlearning in Vision-Language Models

Sudharshan Balaji, Yili Ren, Guangjing Wang, Yimin Chen, Ning Wang

Machine unlearning is widely used to remove hazardous knowledge from large language models. Modern Vision-Language Models (VLMs), however, process both text and visual inputs, raising a fundamental security question: does unlearning in one modality transfer to the other? We prese…

View free PDFSource page
arxivcs.CVcs.AIcs.CLcs.LGcs.MM2026-07-01

ESC: Emotional Self-Correction for Reliable Vision-Language Models

Tien-Huy Nguyen, Minh-Nhat Nguyen, Nguyen Nhat Huy, Hung Viet Nguyen, Huy Nguyen Minh Nhat, Thanh-Huy Nguyen, et al.

Vision-language models (VLMs) have achieved strong performance across diverse multimodal tasks, yet they remain vulnerable to unreliable reasoning. Existing self-correction methods mitigate these issues but typically rely on post-training or carefully engineered feedback, incurri…

View free PDFSource page