CODECRASH: Stress Testing LLM Reasoning under Structural and Semantic Perturbations

Large Language Models (LLMs) have recently showcased strong capabilities in code-related tasks, yet their robustness in code comprehension and reasoning remains underexplored. In this paper, we present CodeCrash, a unified benchmark that evaluates LLM robustness under code structural and textual distraction perturbations, applied to two established benchmarks -- CRUXEval and LiveCodeBench -- across both input and output prediction tasks. We evaluate seventeen LLMs using direct and Chain-of-Thought inference to systematically analyze their robustness, identify primary reasons for performance degradation, and highlight failure modes. Our findings reveal the fragility of LLMs under structural noise and the inherent reliance on natural language cues, highlighting critical robustness issues of LLMs in code execution and understanding. Additionally, we examine three Large Reasoning Models (LRMs) and discover the severe vulnerability of self-reflective reasoning mechanisms that lead to reasoning collapse. CodeCrash provides a principled framework for stress-testing LLMs in code understanding, offering actionable directions for future evaluation and benchmarking. The code of CodeCrash and the robustness leaderboard are publicly available at https://donaldlamnl.github.io/CodeCrash/ .

Fact-or-Fair: A Checklist for Behavioral Testing of AI Models on Fairness-Related Queries

The generation of incorrect images, such as depictions of people of color in Nazi-era uniforms by Gemini, frustrated users and harmed Google's reputation, motivating us to investigate the relationship between accurately reflecting factuality and promoting diversity and equity. In this study, we focus on 19 real-world statistics collected from authoritative sources. Using these statistics, we develop a checklist comprising objective and subjective queries to analyze behavior of large language models (LLMs) and text-to-image (T2I) models. Objective queries assess the models' ability to provide accurate world knowledge. In contrast, the design of subjective queries follows a key principle: statistical or experiential priors should not be overgeneralized to individuals, ensuring that models uphold diversity. These subjective queries are based on three common human cognitive errors that often result in social biases. We propose metrics to assess factuality and fairness, and formally prove the inherent trade-off between these two aspects. Results show that GPT-4o and DALL-E 3 perform notably well among six LLMs and four T2I models. Our code is publicly available at https://github.com/uclanlp/Fact-or-Fair.

How Should I Build A Benchmark?

Various benchmarks have been proposed to assess the performance of large language models (LLMs) in different coding scenarios. We refer to them as code-related benchmarks. However, there are no systematic guidelines by which such a benchmark should be developed to ensure its quality, reliability, and reproducibility. We propose How2Bench, which is comprised of a 55- 55-criteria checklist as a set of guidelines to govern the development of code-related benchmarks comprehensively. Using HOW2BENCH, we profiled 274 benchmarks released within the past decade and found concerning issues. Nearly 70% of the benchmarks did not take measures for data quality assurance; over 10% did not even open source or only partially open source. Many highly cited benchmarks have loopholes, including duplicated samples, incorrect reference codes/tests/prompts, and unremoved sensitive/confidential information. Finally, we conducted a human study involving 49 participants, which revealed significant gaps in awareness of the importance of data quality, reproducibility, and transparency.

MRWeb: An Exploration of Generating Multi-Page Resource-Aware Web Code from UI Designs

Multi-page websites dominate modern web development. However, existing design-to-code methods rely on simplified assumptions, limiting to single-page, self-contained webpages without external resource connection. To address this gap, we introduce the Multi-Page Resource-Aware Webpage (MRWeb) generation task, which transforms UI designs into multi-page, functional web UIs with internal/external navigation, image loading, and backend routing. We propose a novel resource list data structure to track resources, links, and design components. Our study applies existing methods to the MRWeb problem using a newly curated dataset of 500 websites (300 synthetic, 200 real-world). Specifically, we identify the best metric to evaluate the similarity of the web UI, assess the impact of the resource list on MRWeb generation, analyze MLLM limitations, and evaluate the effectiveness of the MRWeb tool in real-world workflows. The results show that resource lists boost navigation functionality from 0% to 66%-80% while facilitating visual similarity. Our proposed metrics and evaluation framework provide new insights into MLLM performance on MRWeb tasks. We release the MRWeb tool, dataset, and evaluation framework to promote further research.

XRZoo: A Large-Scale and Versatile Dataset of Extended Reality (XR) Applications

The rapid advancement of Extended Reality (XR, encompassing AR, MR, and VR) and spatial computing technologies forms a foundational layer for the emerging Metaverse, enabling innovative applications across healthcare, education, manufacturing, and entertainment. However, research in this area is often limited by the lack of large, representative, and highquality application datasets that can support empirical studies and the development of new approaches benefiting XR software processes. In this paper, we introduce XRZoo, a comprehensive and curated dataset of XR applications designed to bridge this gap. XRZoo contains 12,528 free XR applications, spanning nine app stores, across all XR techniques (i.e., AR, MR, and VR) and use cases, with detailed metadata on key aspects such as application descriptions, application categories, release dates, user review numbers, and hardware specifications, etc. By making XRZoo publicly available, we aim to foster reproducible XR software engineering and security research, enable cross-disciplinary investigations, and also support the development of advanced XR systems by providing examples to developers. Our dataset serves as a valuable resource for researchers and practitioners interested in improving the scalability, usability, and effectiveness of XR applications. XRZoo will be released and actively maintained.

C$^2$LEVA: Toward Comprehensive and Contamination-Free Language Model Evaluation

Recent advances in large language models (LLMs) have shown significant promise, yet their evaluation raises concerns, particularly regarding data contamination due to the lack of access to proprietary training data. To address this issue, we present C$^2$LEVA, a comprehensive bilingual benchmark featuring systematic contamination prevention. C$^2$LEVA firstly offers a holistic evaluation encompassing 22 tasks, each targeting a specific application or ability of LLMs, and secondly a trustworthy assessment due to our contamination-free tasks, ensured by a systematic contamination prevention strategy that fully automates test data renewal and enforces data protection during benchmark data release. Our large-scale evaluation of 15 open-source and proprietary models demonstrates the effectiveness of C$^2$LEVA.

On the Shortcut Learning in Multilingual Neural Machine Translation

In this study, we revisit the commonly-cited off-target issue in multilingual neural machine translation (MNMT). By carefully designing experiments on different MNMT scenarios and models, we attribute the off-target issue to the overfitting of the shortcuts of (non-centric, centric) language mappings. Specifically, the learned shortcuts biases MNMT to mistakenly translate non-centric languages into the centric language instead of the expected non-centric language for zero-shot translation. Analyses on learning dynamics show that the shortcut learning generally occurs in the later stage of model training, and multilingual pretraining accelerates and aggravates the shortcut learning. Based on these observations, we propose a simple and effective training strategy to eliminate the shortcuts in MNMT models by leveraging the forgetting nature of model training. The only difference from the standard training is that we remove the training instances that may induce the shortcut learning in the later stage of model training. Without introducing any additional data and computational costs, our approach can consistently and significantly improve the zero-shot translation performance by alleviating the shortcut learning for different MNMT models and benchmarks.

Interaction2Code: How Far Are We From Automatic Interactive Webpage Generation?

Converting webpage design into functional UI code is a critical step for building websites, which can be labor-intensive and time-consuming. To automate this design-to-code transformation process, various automated methods using learning-based networks and multi-modal large language models (MLLMs) have been proposed. However, these studies were merely evaluated on a narrow range of static web pages and ignored dynamic interaction elements, making them less practical for real-world website deployment. To fill in the blank, we present the first systematic investigation of MLLMs in generating interactive webpages. Specifically, we first formulate the Interaction-to-Code task and build the Interaction2Code benchmark that contains 97 unique web pages and 213 distinct interactions, spanning 15 webpage types and 30 interaction categories. We then conduct comprehensive experiments on three state-of-the-art (SOTA) MLLMs using both automatic metrics and human evaluations, thereby summarizing six findings accordingly. Our experimental results highlight the limitations of MLLMs in generating fine-grained interactive features and managing interactions with complex transformations and subtle visual modifications. We further analyze failure cases and their underlying causes, identifying 10 common failure types and assessing their severity. Additionally, our findings reveal three critical influencing factors, i.e., prompts, visual saliency, and textual descriptions, that can enhance the interaction generation performance of MLLMs. Based on these findings, we elicit implications for researchers and developers, providing a foundation for future advancements in this field. Datasets and source code are available at https://github.com/WebPAI/Interaction2Code.

Enhancing Temporal Modeling of Video LLMs via Time Gating

Video Large Language Models (Video LLMs) have achieved impressive performance on video-and-language tasks, such as video question answering. However, most existing Video LLMs neglect temporal information in video data, leading to struggles with temporal-aware video understanding. To address this gap, we propose a Time Gating Video LLM (TG-Vid) designed to enhance temporal modeling through a novel Time Gating module (TG). The TG module employs a time gating mechanism on its sub-modules, comprising gating spatial attention, gating temporal attention, and gating MLP. This architecture enables our model to achieve a robust understanding of temporal information within videos. Extensive evaluation of temporal-sensitive video benchmarks (i.e., MVBench, TempCompass, and NExT-QA) demonstrates that our TG-Vid model significantly outperforms the existing Video LLMs. Further, comprehensive ablation studies validate that the performance gains are attributed to the designs of our TG module. Our code is available at https://github.com/LaVi-Lab/TG-Vid.

Learning to Ask: When LLMs Meet Unclear Instruction

Equipped with the capability to call functions, modern large language models (LLMs) can leverage external tools for addressing a range of tasks unattainable through language skills alone. However, the effective execution of these tools relies heavily not just on the advanced capabilities of LLMs but also on precise user instructions, which often cannot be ensured in the real world. To evaluate the performance of LLMs tool-use under imperfect instructions, we meticulously examine the real-world instructions queried from users, analyze the error patterns, and build a challenging tool-use benchmark called Noisy ToolBench (NoisyToolBench). We find that due to the next-token prediction training objective, LLMs tend to arbitrarily generate the missed argument, which may lead to hallucinations and risks. To address this issue, we propose a novel framework, Ask-when-Needed (AwN), which prompts LLMs to ask questions to users whenever they encounter obstacles due to unclear instructions. Moreover, to reduce the manual labor involved in user-LLM interaction and assess LLMs performance in tool utilization from both accuracy and efficiency perspectives, we design an automated evaluation tool named ToolEvaluator. Our experiments demonstrate that the AwN significantly outperforms existing frameworks for tool learning in the NoisyToolBench. We will release all related code and datasets to support future research.