This paper investigates how microaggressions experienced by female employees can trigger antagonistic work behaviors (AWBs). By drawing on social exchange and reciprocity perspectives, we test a moderated-mediated model of microaggressions toward women. Findings from our vignette experiments (Study 1) conducted with 236 female employees confirmed that individuals exhibit more AWBs in a high-microaggressions condition and vice versa. Additionally, vignette-based Study 2, which used a 2 × 2 factorial design and included 272 female employees, validated the significant interactive effect of psychological ownership (PO) and co-rumination on AWBs. Furthermore, results from a two-wave time-lagged survey (Study 3) conducted with 267 female employees confirmed that microaggressions toward women are significantly related to AWBs. Additionally, we found the significant partial mediating role of PO in the relationship between microaggressions toward women and AWBs. Results regarding the boundary condition showed that high co-rumination at work intensifies the negative relationship between PO and AWBs. We also find evidence of a moderated mediation relationship in our study. These empirical findings add to the existing knowledge on microaggressions toward women and their effects on employee outcomes. Furthermore, these results guide organizations in creating a more diverse workplace where covert forms of discrimination toward women should also be addressed.
Although life satisfaction is a core construct within positive psychology and a widely used indicator of human well-being, research in this area has expanded rapidly and become increasingly fragmented across theoretical perspectives, methodological approaches, and application domains. Despite this growth, no dedicated bibliometric study has systematically mapped the intellectual structure and emerging research directions of life satisfaction research explicitly situated within positive psychology. To address this gap, the present study provides a comprehensive bibliometric analysis to identify current research trends, foundational knowledge structures, and future thematic trajectories. A total of 968 English-language publications indexed in the Web of Science Core Collection from 2000 to December 31, 2025, were analyzed using a science-mapping approach. Two complementary bibliometric techniques were employed: bibliographic coupling to examine the field’s contemporary intellectual structure, and co-word analysis to identify emerging themes and future research directions. Descriptive bibliometric indicators were first used to characterize publication growth and citation patterns, followed by network-based analyses calibrated through threshold testing to ensure robust and interpretable results. The bibliographic coupling analysis revealed four major intellectual clusters: (1) positive psychological interventions and psychosocial resources, (2) conceptual foundations and multidimensional models of well-being, (3) life satisfaction and well-being in children and adolescents, and (4) societal, organizational, and critical perspectives on subjective well-being. The co-word analysis identified four forward-looking thematic clusters: integrative foundations of psychological functioning; psychosocial resources and stress regulation; measurement and validation in youth mental health and education; and intervention-based approaches targeting both well-being enhancement and symptom reduction. Overall, the findings indicate a shift from static, outcome-oriented conceptualizations of life satisfaction toward dynamic, multilevel, and mechanism-focused models embedded within broader psychological, developmental, and contextual systems. By clarifying the intellectual organization and emerging directions of life satisfaction research within positive psychology, this study provides a structured foundation for future theory development, methodological refinement, and evidence-based intervention and policy design.
The World Wide Web spanning blogs, review sites, forums, and social networks—has become a massive hub of information where people share their arguments, opinions, emotions, and perspectives on politics, brands, products, and social events. This user-generated content plays a powerful role in shaping the decisions of readers, influencing advertisers, movie reviews and even swaying politicians. In the field of natural language processing (NLP), analyzing these sentiments is not straightforward. One of the biggest challenges is dealing with incomplete or inconsistently tagged data. This study focuses on key issues such as lexical diversity (the wide variety of words people use), imbalanced datasets (where some sentiments are overrepresented compared to others), and the difficulty of capturing long-distance relationships between words in text. To tackle these challenges, the paper explores a hybrid approach using XLNet, a transformer-based model, combined with Moth-Flame Optimization (MFO). XLNet excels at understanding the contextual relationships between words. Meanwhile, MFO is applied to optimize the model by searching for the best possible weights, ensuring that the most relevant features are used for prediction. The dataset chosen for this study is the IMDB movie reviews, which was first cleaned and preprocessed. XLNet then generated tokens, mask, and segmentation embeddings, while MFO fine-tuned the model to achieve global optimization. This combination significantly improved performance, with the hybrid technique reaching an impressive accuracy of 92.4
Large language models (LLMs) demonstrate strong reasoning abilities via Chain-of-Thought (CoT), but their token-level generation encourages local decisions and lacks global planning, often leading to redundant or inaccurate reasoning. Existing methods, such as tree-based search and reinforcement learning (RL), attempt to address this issue but incur high computational costs and still struggle to produce reliable reasoning trajectories. To address these challenges, we propose Plan-Then-Action Enhanced Reasoning with Group Relative Policy Optimization (PTA-GRPO), a two-stage framework designed to jointly improve high-level planning and fine-grained CoT reasoning. Specifically, in the first stage, a given LLM is responsible for summarizing CoT reasoning into compact high-level guidance, which is then leveraged for supervised fine-tuning. Then, we introduce a guidance-aware reinforcement learning method that jointly optimizes the final output and the quality of guidance, enhancing reasoning effectiveness. We evaluate PTA-GRPO on ten reasoning benchmarks across mathematics and natural sciences, using five diverse base models spanning multiple data modalities. The results show that PTA-GRPO consistently delivers stable and significant improvements across models and tasks, demonstrating strong effectiveness and generalization.
Large language models generate plausible code but cannot verify correctness. Existing multi-agent systems simulate execution or leave verification optional. We introduce execution-grounded verification as a first-class principle: every code change must survive sandboxed execution before propagation. We instantiate this principle in AGENTFORGE, a multi-agent framework where Planner, Coder, Tester, Debugger, and Critic agents coordinate through shared memory and a mandatory Docker sandbox. We formalize software engineering with LLMs as an iterative decision process over repository states, where execution feedback provides a stronger supervision signal than next-token likelihood. AGENTFORGE achieves 40.0% resolution on SWE-BENCH Lite, outperforming single-agent baselines by 26–28 points. Ablations confirm that execution feedback and role decomposition each independently drive performance. The framework is open-source at https://github.com/raja21068/AutoCodeAI.
