OddGridBench: Exposing the Lack of Fine-Grained Visual Discrepancy Sensitivity in Multimodal Large Language Models

2026-03-10 · Tengjin Weng, Wenhao Jiang, Jingyi Wang, Ming Li, Lin Ma, Zhong Ming arxiv

Multimodal large language models (MLLMs) have achieved remarkable performance across a wide range of vision language tasks. However, their ability in low-level visual perception, particularly in detecting fine-grained visual discrepancies, remains underexplored and lacks systematic analysis. In this work, we introduce OddGridBench, a controllable benchmark for evaluating the visual discrepancy sensitivity of MLLMs. OddGridBench comprises over 1,400 grid-based images, where a single element differs from all others by one or multiple visual attributes such as color, size, rotation, or position. Experiments reveal that all evaluated MLLMs, including open-source families such as Qwen3-VL and InternVL3.5, and proprietary systems like Gemini-2.5-Pro and GPT-5, perform far below human levels in visual discrepancy detection. We further propose OddGrid-GRPO, a reinforcement learning framework that integrates curriculum learning and distance-aware reward. By progressively controlling the difficulty of training samples and incorporating spatial proximity constraints into the reward design, OddGrid-GRPO significantly enhances the model's fine-grained visual discrimination ability. We hope OddGridBench and OddGrid-GRPO will lay the groundwork for advancing perceptual grounding and visual discrepancy sensitivity in multimodal intelligence. Code and dataset are available at https://wwwtttjjj.github.io/OddGridBench/.

📄 PDF Abstract BibTeX arXiv:2603.09326

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

VisualNeedle: Benchmarking Active Visual Search in Information-Dense Scenes

2026-05-25 · Jingru Chen, Yiming Liu, Mingtao Chen, Sijie Chen 외 arxiv

Frontier multimodal large language models (MLLMs) have been reported to achieve over 90% accuracy on fine-grained perception benchmarks. However, such scores do not necessarily imply faithful use of visual evidence. Prio…

Subtoken Vision Transformer for Fine-grained Recognition

2026-07-10 · Jie Zhu, Ivy Zhang, Minchul Kim, Xiaoming Liu arxiv

We present Subtoken Vision Transformer (SubViT), a selective image tokenization method for fine-grained visual recognition. Standard Vision Transformers compress each fixed-size patch into a single token, although fine-g…

Fine-Grained Visual Recognition

GradingAttack: Exposing Security Vulnerabilities in LLM Based Educational Grading Agents

2026-02-01 · Xueyi Li, Zhuoneng Zhou, Zitao Liu, Yongdong Wu arxiv

Large language models (LLMs) are increasingly deployed as educational agents for automatic short answer grading (ASAG) in real-world educational environments, significantly boosting assessment efficiency and scalability.…

Adversarial Attack

VDiff-Bench: A Challenging Benchmark for Fine-Grained Image Difference Identification

2026-09-05 · Yixin Wan, Tianle Zheng, Kai-Wei Chang hf

Multimodal Large Language Models (MLLMs) perform strongly on general visual understanding tasks such as visual question answering, yet they often struggle with a basic comparative skill: identifying what has changed betw…

Visual Question Answering

TokenFlow: Rethinking Fine-grained Cross-modal Alignment in Vision-Language Retrieval

2022-09-28 · Xiaohan Zou, Changqiao Wu, Lele Cheng, Zhongyuan Wang

Most existing methods in vision-language retrieval match two modalities by either comparing their global feature vectors which misses sufficient information and lacks interpretability, detecting objects in images or vide…

cross-modal alignment Retrieval Text Retrieval Video-Text Retrieval