False-positive bug reports represent a significant yet underexplored challenge in the development and maintenance of the Linux kernel. They occur when correct system behavior is mistakenly flagged as a defect, consuming developer effort without leading to actual code improvements. Such reports can mislead developers, waste debugging resources, and delay the resolution of real bugs. In this paper, we present the first comprehensive empirical study of false-positive bug reports in the Linux kernel. We manually construct a dataset of 2,006 bug reports comprising 1,509 genuine bugs and 497 false positives collected from Bugzilla and Syzkaller. Our analysis indicates that false positives demand effort comparable to real bugs, often requiring extended discussions and non-trivial closure time. They occur in several components, especially File Systems and Drivers, mainly due to external dependencies and semantic misunderstandings. To address this challenge, we evaluate large language models (LLMs) for automated false-positive bug report mitigation. Among various prompting strategies, retrieval-augmented generation (RAG) performs best, achieving 91

We introduce TableVista, a comprehensive benchmark for evaluating foundation models in multimodal table reasoning under visual and structural complexity. TableVista consists of 3,000 high-quality table reasoning problems, where each instance is expanded into 10 distinct visual variants through our multi-style rendering and transformation pipeline. This process encompasses diverse scenario styles, robustness perturbations, and vision-only configurations, culminating in 30,000 multimodal samples for a multi-dimensional evaluation. We conduct an extensive evaluation of 29 state-of-the-art open-source and proprietary foundation models on TableVista. Through comprehensive quantitative and qualitative analysis, we find that while evaluated models remain largely stable across diverse rendering styles, they exhibit pronounced performance degradation on complex structural layouts and vision-only settings, revealing that current models struggle to maintain reasoning consistency when structural complexity combines with visually integrated presentations. These findings highlight critical gaps in current multimodal capabilities, providing insights for advancing more robust and reliable table understanding models.

Test assertions are critical elements of unit tests, serving as checkpoints to validate expected behavior and ensure software correctness. Numerous techniques have been proposed to automate assertion generation, with recent progress notably driven by large language models (LLMs). Despite the promise, existing approaches such as ChatAssert suffer from modest accuracy, heavy reliance on oversampling, and vulnerability to model randomness due to one-shot prompting. To address these limitations, we propose AssertMate, a novel agent-based assertion generation framework that enhances the quality and reliability of LLM-generated assertions through three key components: (1) actual value construction that identifies assertion targets via static analysis and type-aware heuristics; (2) multi-perspective expected value prediction using code generation, retrieval-augmented generation (RAG), and chain-of-thought (CoT) reasoning agents; and (3) an LLM-as-a-Judge collaboration mechanism to select the most appropriate assertion. Evaluation on the Defects4J benchmark demonstrates that AssertMate significantly outperforms state-of-the-art techniques in compilation success and pass rates, along with substantially higher bug detection capabilities. Integration with EvoSuite further validates AssertMate's practicality, yielding superior mutation coverage and kill counts. Ablation studies reveal that each of the three components makes a significant and complementary contribution to the overall performance. This work affirms the great potential of aggregating diverse perspectives to enhance the effectiveness of LLM-based assertion generation.

As the adoption of LLMs becomes more widespread in software coding ecosystems, a pressing issue has emerged: does the generated code contain social bias and unfairness, such as those related to age, gender, and race? This issue concerns the integrity, fairness, and ethical foundation of software applications that depend on the code generated by these models but are underexplored in the literature. This article presents a novel bias testing framework that is specifically designed for code generation tasks. Based on this framework, we conduct an extensive empirical study on the biases in code generated by five widely studied LLMs (i.e., PALM-2-CodeChat-bison, Claude-instant-1, GPT-3.5-turbo, GPT-4-turbo, and GPT-4). Our findings reveal that biases are prevalent. For example, 13.47% to 49.10% of the codes generated by these LLMs have biased behaviors towards gender. Moreover, we study five bias mitigation prompt strategies that are commonly used in current code generation scenarios, i.e., zero-shot, one-shot, few-shot, and two Chain-of-Thought (CoT) prompts, with and without provided feedback-driven refinement. Our evaluation results illustrate that using direct prompt engineering strategies has limited effectiveness in mitigating bias, but our test execution feedback can help to reduce the ratio of code biases to a large extent (e.g., from 59.88% to 4.79% for GPT-4).

As large language models (LLMs) are increasingly used for long-form generation, reliably evaluating long-form outputs has become a critical challenge. LLM-as-a-judge offers a scalable alternative to human evaluation, yet its reliability in long-form output evaluation remains underexamined: existing meta-evaluation benchmarks focus mainly on short-form outputs. Compared with short-form evaluation, long-form evaluation is not merely a matter of output length; it often requires judges to make more complex document-level assessments of overall organization, task-relevant coverage and depth, cross-section consistency, and scenario-specific quality criteria. In this work, we introduce LongJudgeBench, a comprehensive benchmark for evaluating LLM judges on long-form outputs across diverse real-world scenarios and judging protocols. We systematically evaluate a broad range of LLM judges, covering multiple base models and judging settings. Our results reveal a substantial reliability gap: current LLM judges remain unstable across scenarios, and rubrics or references are helpful but not always sufficient. We hope LongJudgeBench will support future research on more robust, context-aware, and human-aligned LLM-as-a-judge methods. Our code is available at https://github.com/cjj826/LongJudgeBench.

Large Language Models (LLMs) are increasingly deployed across diverse domains, yet their vulnerability to jailbreak attacks, where adversarial inputs bypass safety mechanisms to elicit harmful outputs, poses significant security risks. While prior work has primarily focused on prompt injection attacks, these approaches often require resource-intensive prompt engineering and overlook other critical components, such as chat templates. This paper introduces TEMPLATEFUZZ, a fine-grained fuzzing framework that systematically exposes vulnerabilities in chat templates, a critical yet underexplored attack surface in LLMs. Specifically, TEMPLATEFUZZ (1) designs a series of element-level mutation rules to generate diverse chat template variants, (2) proposes a heuristic search strategy to guide the chat template generation toward the direction of amplifying the attack success rate (ASR) while preserving model accuracy, and (3) integrates an active learning-based strategy to derive a lightweight rule-based oracle for accurate and efficient jailbreak evaluation. Evaluated on twelve open-source LLMs across multiple attack scenarios, TEMPLATEFUZZ achieves an average ASR of 98.2

Mutation testing is a powerful technique for ensuring software quality. However, the presence of equivalent mutants introduces unnecessary costs and biases, limiting its practical effectiveness. Although numerous equivalent mutant detection (EMD) methods have been proposed, they often face distinct challenges: pure-code analysis methods can be limited by their reliance on specific compiler infrastructures, while existing machine-learning approaches remain constrained by scarce training data and limited generalization to unseen mutants. Large language models (LLMs) have recently demonstrated remarkable performance across diverse code-related tasks by better capturing program semantics. Yet their potential for EMD remains largely unexplored, particularly in the multi-lingual context. This paper presents the first comprehensive empirical study on LLMs for EMD, using 3,302 Java and 1,088 C mutant pairs to benchmark against state-of-the-art methods, explore strategy variations, assess efficiency, and evaluate cross-lingual generalization. Experimental results show that LLM-based approaches achieve higher F1-scores than the evaluated traditional methods, with fine-tuned code embedding yielding the highest detection accuracy among the tested strategies. Moreover, LLM-based approaches strike a practical balance between effectiveness and efficiency with inference times comparable to existing machine-learning models. Importantly, fine-tuned LLMs demonstrate measurable generalization across programming languages. These findings establish LLMs as a viable and efficient approach for tackling the longstanding challenge of equivalent mutant detection, offering new directions for advancing mutation testing in practice.

Recent advances in coding agents have shown remarkable progress in software issue resolution. In practice, real-world issues are typically bug fixes or feature requests in which human developers naturally incorporate refactoring as part of the resolution process, resulting in tangled refactoring. Since LLMs are trained on large-scale open-source repositories, coding agents may inherit such behaviors. In this paper, we conduct an empirical study on Multi-SWE-bench, analyzing 3,691 valid patches generated by three agent frameworks with 12 LLMs. We find that coding agents introduce tangled refactorings less frequently (21.43

Garbage Collection (GC) in the Java Virtual Machine (JVM) serves as an automatic memory management mechanism, efficiently reclaiming unused memory space in different production scenarios. To optimize JVM performance, developers typically fine-tune the garbage collector by identifying an optimal set of GC configurations for specific scenarios. Despite the sophisticated design of garbage collectors, they still have the potential for bugs in different settings, and these bugs can result in more severe consequences. Hence, comprehensive testing of these garbage collectors is imperative before their release. Code coverage criteria are typically employed to assess the comprehensiveness of a test suite. However, traditional code coverage metrics, such as branch coverage, are hardly applicable for GC testing due to their inherent concurrency. Additionally, existing JVM testing techniques do not adequately consider the characteristics of garbage collectors, making it difficult to test these garbage collectors sufficiently. In this article, we make the first effort to design coverage criteria against garbage collectors based on the events of GC called Event-Coverage. Its key insight is to measure the diversity of GC executions for testing purposes by assessing the range of GC events these executions cover. Furthermore, we design a new testing method for maximizing Event-Coverage called GCFuzz. GCFuzz conducts an exhaustive investigation of the memory state space of GC and thoroughly explores the memory state under various GC configurations. To enhance GCFuzz's efficiency in achieving higher Event-Coverage, we have further designed a coverage-driven strategy for preserving candidate seed programs and selecting GC configurations. Extensive evaluations demonstrate a positive correlation between Event-Coverage and the bug-revealing efficiency. Moreover, GCFuzz outperforms state-of-the-art techniques in detecting unique GC-related inconsistencies and achieving higher Event-Coverage. Remarkably, GCFuzz has identified 20 previously undetected GC bugs, with 15 of them already confirmed or fixed by developers.

Generative Artificial Intelligence (GenAI) models are achieving remarkable performance in various tasks, including code generation, testing, code review, and program repair. The ability to increase the level of abstraction away from writing code has the potential to change the Human-AI interaction within the integrated development environment (IDE). To explore the impact of GenAI on IDEs, 33 experts from the Software Engineering, Artificial Intelligence, and Human-Computer Interaction domains gathered to discuss challenges and opportunities at Shonan Meeting 222, a four-day intensive research meeting. Four themes emerged as areas of interest for researchers and practitioners.

Code comment generation aims to produce a generic overview of a code snippet, helping developers understand and maintain code. However, generic summaries alone are insufficient to meet the diverse needs of practitioners; for example, developers expect the implementation insights to be presented in an untangled manner, while users seek clear usage instructions. This highlights the necessity of multi-intent comment generation. With the widespread adoption of Large Language Models (LLMs) for code-related tasks, these models have been leveraged to tackle the challenge of multi-intent comment generation. Despite their successes, state-of-the-art LLM-based approaches often struggle to construct correct relationships among intents, code, and comments within a smaller number of demonstration examples. To mitigate this issue, we propose a framework named KUMIC for multi-intent comment generation. Built upon in-context learning, KUMIC leverages Chain-of-Thought (CoT) to optimize knowledge utilization for LLMs to generate intent-specific comments. Specifically, KUMIC first designs a retrieval mechanism to obtain similar demonstration examples, which exhibit high code-comment consistency. Then, KUMIC leverages CoT to guide LLMs to focus on statements facilitating the derivation of code comments aligned with specific intents. In this context, KUMIC constructs a mapping knowledge chain, linking code to intent-specific statements to comments, which enables LLMs to follow similar reasoning steps when generating the desired comments. We conduct extensive experiments to evaluate KUMIC, and the results demonstrate that KUMIC outperforms state-of-the-art baselines by 14.49%, 22.41%, 20.72%, and 12.94% in terms of BLEU, METEOR, ROUGE-L, and SBERT, respectively.

Automated skill evolution enables Large Language Model (LLM) agents to continuously improve without expensive retraining. However, existing approaches typically treat skill evolution as a sequence of local updates, overlooking relationships among skills and often producing overfitted skill updates that fail to generalize across tasks. We propose GSE, a globalized skill evolution framework that jointly optimizes skill compatibility and skill generalization. To preserve consistency across the skill bank, GSE maintains a Skill Relation Graph (SRG) that explicitly models and co-evolves inter-skill relationships. To improve generalization, GSE performs cluster-based skill consolidation to abstract reusable capabilities from local updates and employs replay-driven verification to prevent overfitting and behavioral regressions. We evaluate GSE on two representative software engineering tasks: bug-revealing test generation and false-positive bug report filtering. Across two state-of-the-art coding agents, OpenHands and mini-SWE-agent, GSE consistently achieves the best precision, recall, and F1-score. Compared with existing evolution techniques, GSE improves precision and recall by 6.1

Processor functional verification plays a crucial role in ensuring the quality of processor designs. Traditional techniques like Constrained Random Verification (CRV) struggle to achieve high functional coverage due to the vast instruction space of processors. While LLM-based techniques show potential, merely instructing LLMs has notable limitations, especially when addressing functional points that require deep semantic understanding. To tackle these challenges, we propose a novel technique, FLAME, which specially designs and integrates Retrieval-Augmented Generation (RAG), Chain-of-Thought (CoT), and functional-coverage-guided feedback strategies. This technique establishes semantic mappings between functional points and instructions, enabling the iterative generation of valid and effective test cases. Evaluation on four widely-used open-source processor designs shows that FLAME outperforms the typical and state-of-the-art baselines in functional coverage improvement with an average of 34.25%∼220% while drastically reducing the time required to achieve the same functional coverage by up to 86.13%. Moreover, ablation analysis highlights the vital role of each component in the framework’s overall effectiveness. This work demonstrates the superiority of our LLM-based technique FLAME in enhancing processor functional verification.

Large Language Model (LLM) agents are increasingly deployed in practice across a wide range of autonomous applications. Yet current safety mechanisms for LLM agents focus almost exclusively on preventing failures in advance, providing limited capabilities for responding to, containing, or recovering from incidents after they inevitably arise. In this work, we introduce AIR, the first incident response framework for LLM agent systems. AIR defines a domain-specific language for managing the incident response lifecycle autonomously in LLM agent systems, and integrates it into the agent's execution loop to (1) detect incidents via semantic checks grounded in the current environment state and recent context, (2) guide the agent to execute containment and recovery actions via its tools, and (3) synthesize guardrail rules during eradication to block similar incidents in future executions. We evaluate AIR on three representative agent types. Results show that AIR achieves detection, remediation, and eradication success rates all exceeding 90%. Extensive experiments further confirm the necessity of AIR's key design components, show the timeliness and moderate overhead of AIR, and demonstrate that LLM-generated rules can approach the effectiveness of developer-authored rules across domains. These results show that incident response is both feasible and essential as a first-class mechanism for improving agent safety.

Code generation, which aims to automatically generate source code from given programming requirements, has the potential to substantially improve software development efficiency. With the rapid advancement of large language models (LLMs), LLM-based code generation has attracted widespread attention from both academia and industry. However, as programming requirements become increasingly complex, existing LLMs still exhibit notable performance limitations. To address this challenge, recent studies have proposed training-based curriculum reinforcement learning (CRL) strategies to improve LLM code generation performance. Despite their effectiveness, existing CRL approaches suffer from several limitations, including misaligned requirement difficulty perception, the absence of requirement difficulty optimization, and suboptimal curriculum sampling strategies. In CRL-based code generation, programming requirements serve as the sole input to the model, making their quality and difficulty critical to training effectiveness. Motivated by insights from software requirements engineering, we propose RECRL, a novel requirement-aware curriculum reinforcement learning framework for enhancing LLM-based code generation. RECRL automatically perceives model-specific requirement difficulty, optimizes challenging requirements to improve training data utilization, and employs an adaptive curriculum sampling strategy to construct training batches with smoothly varying difficulty. Extensive experiments on five state-of-the-art LLMs across five widely-used code generation benchmarks by comparing with five state-of-the-art baselines, demonstrate the significant effectiveness of RECRL. For example, RECRL achieves an average Pass@1 improvement of 1.23

Automated unit test generation has recently benefited from advances in large language models (LLMs), yet our industrial deployments reveal a persistent gap between promising research results and practical usability. In real-world projects with complex frameworks and cross-file dependencies, LLM-generated tests frequently fail to compile, require costly manual repair, or provide unstable coverage improvements. This paper reports our experience in designing, deploying, and evaluating CATGen, a context-aware workflow for LLM-based unit test generation, informed by repeated industrial failures and refinements. Rather than relying on LLMs to infer incomplete project context, we found that compilation robustness critically depends on making project-level dependencies explicit, stabilizing test class scaffolding, and replacing iterative LLM-based repair with lightweight static analysis. These experience-driven insights shaped CATGen's multi-stage design, which combines structured context retrieval, deterministic test skeleton construction, and program analysis-based post-processing. We evaluate CATGen on real-world complex focal methods from proprietary industrial projects and additionally on the Defects4J benchmark to assess generalizability. Across both settings, CATGen substantially improves compilation success and structural coverage while significantly reducing generation time and token consumption compared to existing LLM-based approaches. Our results demonstrate that reliable LLM-based unit test generation in practice depends less on prompt engineering alone and more on systematic engineering support grounded in real-world development constraints.

As software systems grow in complexity, particularly in safety-critical domains, detecting defects early is crucial to prevent severe failures. Among existing automated defect detection methods, rule-based systems excel at catching well-understood mistakes but are hard to generalize beyond their predefined rules. To address this challenge, we introduce DslGen, a novel approach that automatically extracts high-quality defect patterns from singular defect repair examples by leveraging large language models (LLMs). Specifically, it first infers the semantic intent of a defect repair for capturing the defect’s essential characteristics via LLM-powered analysis, then generates precise constraint rules to define the defect pattern. Unlike prior methods, DslGen does not require numerous examples or rigid abstraction schemes, enabling robust generalization even from a single fix. To evaluate the effectiveness of DslGen, we evaluated it in two distinct scenarios: (1) controlled experiments and (2) a simulated real-world defect detection pipeline. In the controlled setting, results show that DslGen outperforms all baselines, achieving a precision up to 81.1% and a recall of over 73.0%. In the simulated scenario, DslGen also performed best with an average improvement of 425.1% over state-of-the-art competitors (including GPT-4o). A user study with 15 experienced developers further confirmed DslGen’s practicality, with participants praising the readability and generalizability of its generated DSLs and expressing strong adoption intent.

Fairness testing aims at mitigating unintended discrimination in the decision-making process of data-driven AI systems. Individual discrimination may occur when an AI model makes different decisions for two distinct individuals who are distinguishable solely according to protected attributes, such as age and race. Such instances reveal biased AI behaviour, and are called Individual Discriminatory Instances (IDIs). In this paper, we propose an approach for the selection of the initial seeds to generate IDIs for fairness testing. Previous studies mainly used random initial seeds to this end. However this phase is crucial, as these seeds are the basis of the follow-up IDIs generation. We dubbed our proposed seed selection approach I&D. It generates a large number of initial IDIs exhibiting a great diversity, aiming at improving the overall performance of fairness testing. Our empirical study reveals that I&D is able to produce a larger number of IDIs with respect to four state-of-the-art IDI generation approaches, generating 1.86X more IDIs on average. When using the IDIs generated with I&D for retraining a machine learning model, the percentage of IDIs in the input space \(\mathbb{I}\) is decreased by 24.9% on average, implying that I&D is effective for improving the model’s fairness.