arXiv Computation and Language

Modality Fault Lines: Structural Corruptions Reveal Fragile Omni-Modal Reasoning

arXiv Computation and Language
Aug 25

Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models

Omni‑SafetyBench is a new benchmark designed to evaluate the safety of Omni‑Modal Large Language Models (OLLMs) that process visual, auditory, and textual data. It contains 23,328 test instances across 24 modality variations derived from 972 seed samples, and introduces metrics such as Safety‑score (based on Conditional Attack Success Rate and Conditional Refusal Rate) and Cross‑Modal Safety Consistency score. Evaluation of 11 state‑of‑the‑art OLLMs shows severe vulnerabilities, with only three models achieving a Safety‑score above 0.6 and safety degrading sharply for audio‑visual inputs, underscoring the need for improved safety alignment methods.

By Leyi Pan, Zheyu Fu, Yunpeng Zhai, Shuchang Tao, Sheng Guan, Shiyu Huang, Lingzhe Zhang, Zhaoyang Liu, Bolin Ding, Felix Henry, Aiwei Liu, Lijie Wen
arXiv AI
Aug 26

OmniJudge or OmniBias? Diagnosing Multimodal Judges through Balanced, Decoupled Lenses

The paper introduces D3-Omni, a balanced and decoupled benchmark designed to diagnose fine‑grained multimodal understanding in OmniJudges that evaluate text‑to‑image, text‑to‑video, and text‑to‑speech generation. D3-Omni covers 53 orthogonal binary dimensions across 10,671 samples, using fixed positive seeds and controlled prompt rewriting to generate negatives, thereby ensuring each error can be attributed to a single capability. The benchmark’s dual‑balanced, decoupled, and dynamic design achieves near 1:1 per‑dimension parity and a uniform total‑score distribution, revealing that strong OmniJudges often miss modality‑related failures and treat distinct attributes as a single decision, masking systematic blind spots.

By Guangzheng Hu, Ziyue Jiang, Weixu Qiao, Lixin Zhang, Jianye Kang, Yuru Wu, Rong Bao, Niantong Li, Wei Wang, Ziyi Cheng, Xinfa Zhu, HangRui Hu, Ting He, Bing Zhao, Lin Qu, Hu Wei, Jin Xu
arXiv Computation and Language
Aug 28

Said Aloud, Read Different: Cross-Modal Instability in Multimodal Models

The paper introduces a new benchmark called Speech-Augmented Visually Grounded Contrastive Triplet Benchmark, comprising 10,150 images from 18 MENA countries, each paired with a supported statement and two plausible but unsupported alternatives. It defines contrastive instability as the rate at which multimodal models fail to resolve all statements within a triplet, distinguishing fragmented reasoning from complete failure. Experiments on recent multimodal models show that shifts in modality (text vs. speech) and language (English vs. Arabic) lead to significant triplet-level inconsistencies, especially when speech is used, which are not fully reflected by overall accuracy metrics.

By Basel Mousi, Fahim Dalvi, Shammur Chowdhury, Firoj Alam, Nadir Durrani
arXiv Computation and Language
2d ago

OmniHallu: Unified Hallucination Detection for Cross-Modal Comprehension and Generation in Multimodal Large Language Models

OmniHallu is a unified framework for detecting hallucinations in multimodal large language models across both comprehension and generation tasks involving image, video, and audio modalities. It introduces OmniHallu-Bench, a 10,000-sample benchmark with claim-level human annotations for six cross-modal tasks (I2T, V2T, A2T, T2I, T2V, T2A). The system uses a multi‑agent architecture that decomposes outputs into atomic claims, verifies them with modality‑specific experts, and aggregates evidence through structured reasoning, while a preference‑optimized verifier reduces expert calls by 66% with minimal performance loss.

By Jianjiang Yang, Peihang Li, Shanqing Xu, Mengchen Qian, Lu Zhang, Meng Luo