It is found that the original safety alignment of the underlying LLM is not consistently preserved in current unified models, and open-source UMLMs exhibit much lower safety performance than multimodal large models specialized for either generation or understanding tasks, particularly on the generation side.
Abstract
Unified Multimodal Large Models (UMLMs) integrate understanding and generation capabilities within a single architecture. While unified architectures expand multimodal capabilities, their safety implications remain important yet underexplored. Existing safety benchmarks predominantly focus on isolated understanding or generation tasks, failing to evaluate the holistic safety of UMLMs when handling diverse tasks under a unified framework. To address this, we introduce Uni-SafeBench, a comprehensive benchmark featuring a taxonomy of six major safety categories across seven task types. To ensure rigorous assessment, we develop Uni-Judger, a framework that effectively decouples contextual safety from intrinsic safety. Based on comprehensive evaluations across Uni-SafeBench, we find that the original safety alignment of the underlying LLM is not consistently preserved in current unified models. Moreover, open-source UMLMs exhibit much lower safety performance than multimodal large models specialized for either generation or understanding tasks, particularly on the generation side.
While Large Language Models (LLMs) have demonstrated remarkable capabilities in code generation, their effectiveness in unit testing is often constrained by insufficient context regarding external dependencies. This limitation is particularly pronounced in industrial settings, where proprietary code remains opaque to the model. To address this challenge, we present a systematic empirical study of multiple strategies for context enrichment and optimization in LLM‐based unit test generation, conducted on seven diverse projects (three open‐source and four proprietary industrial systems), encompassing 261 distinct methods. By evaluating seven implementations (ranging from basic prompts to optimized context reduction strategies) across 10 independent runs, we analysed a total of 28,710 test suites. Our results demonstrate that combining prompt engineering with external dependency retrieval achieves an average branch coverage increase of 11.52 percentage points on industrial software over the baseline, with statistically significant improvements across all competing implementations. Beyond coverage, richer context substantially reduces generation‐repair iterations, cutting median execution time by 51.3% in industrial projects. We further show that reducing external dependencies to method signatures alone decreases input token consumption by up to 46.6% (25.4% in industrial projects) while fully preserving the coverage and efficiency gains of the complete retrieval approach. To confirm that these benefits are not tied to a specific model, we replicate the core comparison across three LLM backends from different families, obtaining a consistent, statistically significant coverage improvement on industrial code in every case. These findings establish this optimized context strategy as a cost‐effective solution for scalable, industrial‐grade automated test generation.
Javier Ferrer, Francisco Chicano· Expert systems· 0 citations
ReFrame is a training-free multimodal input reframing framework where two agents share a lightweight locally deployed MLLM: the evidence-generation agent constructs complementary risk and utility evidence, and the rewrite-and-routing agent converts it into a safe proxy prompt and image-routing decision before calling the downstream MLLM, without modifying it or accessing its internal information.
Wenzheng Jiang, Xuankun Rong, Yuanzhao Zhai et al.· 0 citations
In high-risk industrial settings, leveraging large language models (LLMs) for automated accident analysis and generating safety reports has emerged as an efficient workflow. However, this approach is fundamentally constrained by the models’ inherent knowledge limitations, frequently resulting in analyses that lack domain-specific understanding and regulatory alignment.
To tackle this issue, we introduce SafeRAG, a multistage retrieval-augmented framework for safety report generation. Specifically, the framework uses an entity-centric approach that prompts the LLMs to internally generate domain-specific knowledge. Concurrently, it performs a hierarchical retrieval of external regulations relevant to the accident at topic, concept, and context levels. To obtain well-structured reports, we leverage prompt engineering, integrating internal and external knowledge. Furthermore, a domain-expert persona is also assigned to help LLMs analyze accidents from a specific perspective. To evaluate our approach, we construct a data set from 10,818 accident-description/report pairs collected from real-world industry reports. Experiments show that SafeRAG substantially outperforms baseline LLMs on metrics that include bidirectional encoder representations from transformers (BERTScore) and bidirectional auto-regressive transformers (BARTScore), demonstrating the effectiveness of our approach.
Qi Cao, Chiyu Zhang, Junxiong Ning et al.· SPE Journal· 0 citations
Multimodal large language models (MLLMs) are increasingly expected to automate visualization development by generating code directly from visual designs. However, existing evaluations mainly focus on single-chart generation and overlook coordinated multi-view interface construction, which requires joint reasoning about data semantics, view coordination, and interaction logic. Consequently, MLLM capabilities in this setting remain underexplored, and the field lacks a dedicated benchmark for systematic assessment. We introduce MV-Bench, a benchmark for evaluating MLLMs on coordinated multi-view interface construction. Instead of relying on incomplete or inconsistent open-source implementations, we use Tableau workbook files as ground truth because they explicitly encode data bindings, visual mappings, and interactions. We develop a multi-stage pipeline that converts these specifications into executable web interfaces through structured intermediate representations. The benchmark contains 92 base interfaces and 1,048 verified instances created by recombining chart types, datasets, and interaction patterns. Each instance includes executable code, a rendered interface, a dataset, and interaction annotations. We evaluate five state-of-the-art MLLMs in a single-pass setting using metrics for visual fidelity, data binding correctness, and interaction completeness. The strongest model achieves 75.45 percent accuracy in visual layout reproduction, but only 21.71 percent in data binding and 11.68 percent in interaction completeness. These results show that current MLLMs can reproduce visual appearance but remain limited in generating the data semantics and interactive logic required by coordinated multi-view interfaces. Iterative refinement improves code executability but does not substantially reduce the gap in data binding and interaction generation.
Yue Zhao, Hongxu Liu, Feiyu Wang et al.· 0 citations
Eagle, a comprehensive benchmarking framework tailored for evaluating OpsLLMs, delivers a deployable foundation for advancing large-model applications in AIOps and open-source the framework and dataset to foster community adoption and reproducibility.
Yuhe Liu, Changhua Pei, Hang Wang et al.· SIGSOFT FSE Companion· 0 citations
