The recent shift in Generative AI (GenAI) applications from cloud-only environments to end-user devices introduces new challenges in resource management, system efficiency, and user experience. This paper presents ConsumerBench, a comprehensive benchmarking framework designed to evaluate the system efficiency and response time of GenAI models running on end-user devices. Unlike existing benchmarks that assume exclusive model access on dedicated GPUs, ConsumerBench simulates realistic multi-application scenarios executing concurrently on constrained hardware. Furthermore, ConsumerBench supports customizable workflows that simulate complex tasks requiring coordination among multiple applications. ConsumerBench captures both application-level metrics, including latency and Service Level Objective (SLO) attainment, and system-level metrics like CPU/GPU utilization and memory bandwidth. Through extensive experiments, ConsumerBench reveals inefficiencies in resource sharing, unfair scheduling under greedy allocation, and performance pitfalls of static model server configurations. The paper also provides practical insights for model developers and system designers, highlighting the benefits of custom kernels tailored to consumer-grade GPU architectures and the value of implementing SLO-aware scheduling strategies.

In this paper, we propose DuaShepherd, a novel reward modeling framework that integrates two complementary reward signals, correctness and potential, to enhance the mathematical reasoning capabilities of Large Language Models (LLMs). While correctness-based signals emphasize identification of stepwise errors, potential-based signals focus on the likelihood of reaching the correct final answer. We developed an automated pipeline for constructing large-scale reward modeling dataset with both signals. A unified, multi-head architecture was explored to train the two reward models in a multi-task setup, demonstrating benefits from learning both correctness and potential in parallel. By combining these two signals into a compound probability, our model achieves consistent performance improvements across multiple benchmarks. Empirical evaluations on MATH500 and ProcessBench confirm that this combined reward significantly outperforms models trained on either reward type alone, achieving state-of-the-art performance under comparable resource constraints.

Software development is currently under a paradigm shift in which Artificial Intelligence (AI) - in particular Generative AI [6] - has taken an increasingly central role in assisting developers in their software creation activities. This is in essence a new form of software reuse in which collections of previously created software artifacts form the basis for generating new ones. Unlike in the past when developers were manually searching for pre-existing software components from libraries and code repositories such as Github, Node Package Manager (NPM) or the Python Package Index (PyPI), in the new model developers are requesting AI-driven assistants to generate suitable pieces of code for them. These generated artifacts can range from small code snippets and module fragments to comprehensive application skeletons or in some cases fully functional applications or even complete end-to-end systems. This new generative approach to software reuse has resulted in a considerable mental model change for developers.

While Large Language Models (LLMs) demonstrate impressive reasoning capabilities, growing evidence suggests much of their success stems from memorized answer-reasoning patterns rather than genuine inference. In this work, we investigate a central question: are LLMs primarily anchored to final answers or to the textual pattern of reasoning chains? We propose a five-level answer-visibility prompt framework that systematically manipulates answer cues and probes model behavior through indirect, behavioral analysis. Experiments across state-of-the-art LLMs reveal a strong and consistent reliance on explicit answers. The performance drops by 26.90\% when answer cues are masked, even with complete reasoning chains. These findings suggest that much of the reasoning exhibited by LLMs may reflect post-hoc rationalization rather than true inference, calling into question their inferential depth. Our study uncovers the answer-anchoring phenomenon with rigorous empirical validation and underscores the need for a more nuanced understanding of what constitutes reasoning in LLMs.

AI systems will soon have to navigate human environments and make decisions that affect people and other AI agents whose goals and values diverge. Contractualist alignment proposes grounding those decisions in agreements that diverse stakeholders would endorse under the right conditions, yet securing such agreement at scale remains costly and slow -- even for advanced AI. We therefore propose Resource-Rational Contractualism (RRC): a framework where AI systems approximate the agreements rational parties would form by drawing on a toolbox of normatively-grounded, cognitively-inspired heuristics that trade effort for accuracy. An RRC-aligned agent would not only operate efficiently, but also be equipped to dynamically adapt to and interpret the ever-changing human social world.

Many researchers are working to address the worldwide mental health crisis by developing therapeutic technologies that increase the accessibility of care, including leveraging large language model (LLM) capabilities in chatbots and socially assistive robots (SARs) used for therapeutic applications. Yet, the effects of these technologies over time remain unexplored. In this study, we use a factorial design to assess the impact of embodiment and time spent engaging in therapeutic exercises on participant disclosures. We assessed transcripts gathered from a two-week study in which 26 university student participants completed daily interactive Cognitive Behavioral Therapy (CBT) exercises in their residences using either an LLM-powered SAR or a disembodied chatbot. We evaluated the levels of active engagement and high intimacy of their disclosures (opinions, judgments, and emotions) during each session and over time. Our findings show significant interactions between time and embodiment for both outcome measures: participant engagement and intimacy increased over time in the physical robot condition, while both measures decreased in the chatbot condition.

Ultra-long generation by large language models (LLMs) is a widely demanded scenario, yet it remains a significant challenge due to their maximum generation length limit and overall quality degradation as sequence length increases. Previous approaches, exemplified by LongWriter, typically rely on ''teaching'', which involves supervised fine-tuning (SFT) on synthetic long-form outputs. However, this strategy heavily depends on synthetic SFT data, which is difficult and costly to construct, often lacks coherence and consistency, and tends to be overly artificial and structurally monotonous. In this work, we propose an incentivization-based approach that, starting entirely from scratch and without relying on any annotated or synthetic data, leverages reinforcement learning (RL) to foster the emergence of ultra-long, high-quality text generation capabilities in LLMs. We perform RL training starting from a base model, similar to R1-Zero, guiding it to engage in reasoning that facilitates planning and refinement during the writing process. To support this, we employ specialized reward models that steer the LLM towards improved length control, writing quality, and structural formatting. Experimental evaluations show that our LongWriter-Zero model, trained from Qwen2.5-32B, consistently outperforms traditional SFT methods on long-form writing tasks, achieving state-of-the-art results across all metrics on WritingBench and Arena-Write, and even surpassing 100B+ models such as DeepSeek R1 and Qwen3-235B. We open-source our data and model checkpoints under https://huggingface.co/THU-KEG/LongWriter-Zero-32B

Large Language Models face significant challenges in maintaining coherent interactions over extended dialogues due to their limited contextual memory. This limitation often leads to fragmented exchanges and reduced relevance in responses, diminishing user experience. To address these issues, we propose a memory-augmented architecture that dynamically retrieves, updates, and prunes relevant information from past interactions, ensuring effective long-term context handling. Experimental results demonstrate that our solution significantly improves contextual coherence, reduces memory overhead, and enhances response quality, showcasing its potential for real-time applications in interactive systems.

The rapid expansion of information from diverse sources has heightened the need for effective automatic text summarization, which condenses documents into shorter, coherent texts. Summarization methods generally fall into two categories: extractive, which selects key segments from the original text, and abstractive, which generates summaries by rephrasing the content coherently. Large language models have advanced the field of abstractive summarization, but they are resource-intensive and face significant challenges in retaining key information across lengthy documents, which we call being "lost in the middle". To address these issues, we propose a hybrid summarization approach that combines extractive and abstractive techniques. Our method splits the document into smaller text chunks, clusters their vector embeddings, generates a summary for each cluster that represents a key idea in the document, and constructs the final summary by relying on a Markov chain graph when selecting the semantic order of ideas.

Large Language Models (LLMs) in mental healthcare risk propagating biases that reinforce stigma and harm marginalized groups. While previous research identified concerning trends, systematic methods for detecting intersectional biases remain limited. This work introduces a multi-hop question answering (MHQA) framework to explore LLM response biases in mental health discourse. We analyze content from the Interpretable Mental Health Instruction (IMHI) dataset across symptom presentation, coping mechanisms, and treatment approaches. Using systematic tagging across age, race, gender, and socioeconomic status, we investigate bias patterns at demographic intersections. We evaluate four LLMs: Claude 3.5 Sonnet, Jamba 1.6, Gemma 3, and Llama 4, revealing systematic disparities across sentiment, demographics, and mental health conditions. Our MHQA approach demonstrates superior detection compared to conventional methods, identifying amplification points where biases magnify through sequential reasoning. We implement two debiasing techniques: Roleplay Simulation and Explicit Bias Reduction, achieving 66-94% bias reductions through few-shot prompting with BBQ dataset examples. These findings highlight critical areas where LLMs reproduce mental healthcare biases, providing actionable insights for equitable AI development.