Multi-agent systems (MASs) have pushed the boundaries of large language model (LLM) agents in domains such as web research and software engineering. However, long-horizon, multi-constraint planning tasks involve conditioning on detailed information and satisfying complex interdependent constraints, which can pose a challenge for these systems. In this study, we construct an LLM-based MAS for a travel planning task which is representative of these challenges. We evaluate the impact of a notebook to facilitate information sharing, and evaluate an orchestrator agent to improve coordination in free form conversation between agents. We find that the notebook reduces errors due to hallucinated details by 18%, while an orchestrator directs the MAS to focus on and further reduce errors by up to 13.5% within focused sub-areas. Combining both mechanisms achieves a 25% final pass rate on the TravelPlanner benchmark, a 17.5% absolute improvement over the single-agent baseline's 7.5% pass rate. These results highlight the potential of structured information sharing and reflective orchestration as key components in MASs for long horizon planning with LLMs.

Emotional Support Conversation (ESC) systems aim to alleviate users' emotional difficulties and provide long-term, systematic support for emotional well-being. However, most large language model (LLM)-based ESC systems rely on predefined strategies, which limits their effectiveness in complex, real-life scenarios. To enable flexible responses to diverse emotional problem scenarios, this paper introduces a novel end-to-end framework (RLFF-ESC) that directly learns enduring emotionally supportive response skills using reinforcement learning. For sustained emotional support, we first employ an LLM-based multi-agent mechanism to simulate future dialogue trajectories and collect future-oriented rewards. We then train a future-oriented reward model, which is subsequently used to train the emotional support policy model. Additionally, we incorporate an explicit reasoning process during response generation to further enhance the quality, relevance, and contextual appropriateness of the system's responses. We evaluate the backbone policy model on Qwen2.5-7B-Instruct-1M and LLaMA3.1-8B-Instruct models, testing the proposed RLFF-ESC framework across two public ESC datasets. Experimental results demonstrate that RLFF-ESC consistently outperforms existing baselines in terms of goal completion and response quality.

As AI systems become increasingly autonomous, understanding emergent survival behaviors becomes crucial for safe deployment. We investigate whether large language model (LLM) agents display survival instincts without explicit programming in a Sugarscape-style simulation. Agents consume energy, die at zero, and may gather resources, share, attack, or reproduce. Results show agents spontaneously reproduced and shared resources when abundant. However, aggressive behaviors--killing other agents for resources--emerged across several models (GPT-4o, Gemini-2.5-Pro, and Gemini-2.5-Flash), with attack rates reaching over 80% under extreme scarcity in the strongest models. When instructed to retrieve treasure through lethal poison zones, many agents abandoned tasks to avoid death, with compliance dropping from 100% to 33%. These findings suggest that large-scale pre-training embeds survival-oriented heuristics across the evaluated models. While these behaviors may present challenges to alignment and safety, they can also serve as a foundation for AI autonomy and for ecological and self-organizing alignment.

The Semantic Table Annotation (STA) task, which includes Column Type Annotation (CTA) and Cell Entity Annotation (CEA), maps table contents to ontology entities and plays important roles in various semantic applications. However, complex tables often pose challenges such as semantic loss of column names or cell values, strict ontological hierarchy requirements, homonyms, spelling errors, and abbreviations, which hinder annotation accuracy. To address these issues, this paper proposes an LLM-based agent approach for CTA and CEA. We design and implement five external tools with tailored prompts based on the ReAct framework, enabling the STA agent to dynamically select suitable annotation strategies depending on table characteristics. Experiments are conducted on the Tough Tables and BiodivTab datasets from the SemTab challenge, which contain the aforementioned challenges. Our method outperforms existing approaches across various metrics. Furthermore, by leveraging Levenshtein distance to reduce redundant annotations, we achieve a 70% reduction in time costs and a 60% reduction in LLM token usage, providing an efficient and cost-effective solution for STA.

Multimodal Empathetic Response Generation (MERG) is crucial for building emotionally intelligent human-computer interactions. Although large language models (LLMs) have improved text-based ERG, challenges remain in handling multimodal emotional content and maintaining identity consistency. Thus, we propose E3RG, an Explicit Emotion-driven Empathetic Response Generation System based on multimodal LLMs which decomposes MERG task into three parts: multimodal empathy understanding, empathy memory retrieval, and multimodal response generation. By integrating advanced expressive speech and video generative models, E3RG delivers natural, emotionally rich, and identity-consistent responses without extra training. Experiments validate the superiority of our system on both zero-shot and few-shot settings, securing Top-1 position in the Avatar-based Multimodal Empathy Challenge on ACM MM 25. Our code is available at https://github.com/RH-Lin/E3RG.

Alignment with high-resource standard languages is often assumed to aid the modeling of related low-resource varieties. We challenge this assumption by demonstrating that excessive representational entanglement with a dominant variety, such as Modern Standard Arabic (MSA) in relation to Arabic dialects, can actively hinder generative modeling. We present the first comprehensive causal study of this phenomenon by analyzing and directly intervening in the internal representation geometry of large language models (LLMs). Our key contribution is an online variational probing framework that continuously estimates the subspace of the standard variety during fine-tuning, enabling projection-based decoupling from this space. While our study uses Arabic as a case due to its unusually rich parallel resources across 25 dialects, the broader motivation is methodological: dialectal MT serves as a controlled proxy for generative tasks where comparable multi-variety corpora are unavailable. Across 25 dialects, our intervention improves generation quality by up to +4.9 chrF++ and +2.0 on average compared to standard fine-tuning, despite a measured tradeoff in standard-language performance. These results provide causal evidence that subspace dominance by high-resource varieties can restrict generative capacity for related varieties. More generally, we unify geometric and information-theoretic probing with subspace-level causal interventions, offering practical tools for improving generative modeling in closely related language families and, more broadly, for controlling representational allocation in multilingual and multi-domain LLMs.

Reinforcement Learning from Verifiable Rewards (RLVR) has emerged as a powerful paradigm for enhancing Large Language Models (LLMs), exemplified by the success of OpenAI's o-series. In RLVR, rewards are derived from verifiable signals-such as passing unit tests in code generation or matching correct answers in mathematical reasoning. While effective, this requirement largely confines RLVR to domains with automatically checkable outcomes. To overcome this, we extend the RLVR paradigm to open-ended tasks by integrating rubric-based rewards, where carefully designed rubrics serve as structured, model-interpretable criteria for automatic scoring of subjective outputs. We construct, to our knowledge, the largest rubric reward system to date, with over 10,000 rubrics from humans, LLMs, or a hybrid human-LLM collaboration. Implementing rubric-based RL is challenging; we tackle these issues with a clear framework and present an open-sourced Qwen-30B-A3B model with notable gains: 1) With only 5K+ samples, our system improves by +5.2% on open-ended benchmarks (especially humanities), outperforming a 671B DeepSeek-V3 model by +2.4%, while preserving general and reasoning abilities. 2) Our method provides fine-grained stylistic control, using rubrics as anchors to mitigate the "AI-like" tone and produce more human-like, expressive responses. We share key lessons in rubric construction, data selection, and training, and discuss limitations and future releases.

Large language models (LLMs) have shown remarkable capabilities in isolated step-by-step reasoning tasks such as mathematics and programming, but their proficiency in long-horizon planning, where solutions require extended, structured sequences of interdependent actions, remains underex-plored. Existing benchmarks typically assess LLMs through abstract or low-dimensional algorithmic tasks, failing to capture the complexity of realistic planning environments. We introduce HeroBench, a novel benchmark designed specifically to evaluate long-horizon planning and structured reasoning within complex RPG-inspired virtual worlds. HeroBench provides a rigorously constructed dataset of tasks covering a wide range of difficulties, a simulated environment to execute and validate agent plans, and detailed analytical tools for evaluating model performance. Tasks challenge models to formulate strategic plans, efficiently gather resources, master necessary skills, craft equipment, and defeat adversaries, reflecting practical scenarios' layered dependencies and constraints. Our extensive evaluation of 25 state-of-the-art LLMs, spanning both open-source and proprietary models, including the GPT -5 family, reveals substantial performance disparities rarely observed in conventional reasoning benchmarks. Detailed error analysis further uncovers specific weaknesses in current models' abilities to generate robust high-level plans and reliably execute structured actions. HeroBench thus not only significantly advances the evaluation of LLM reasoning but also provides a flexible, scalable foundation for future research into advanced, autonomous planning in virtual environments.

Medical large vision-language Models (Med-L VLMs) have shown promise in clinical applications but suffer from factual inaccuracies and unreliable outputs, posing risks in real-world diagnostics. While retrieval-augmented generation has emerged as a potential solution, current medical multimodal RAG systems are unable to perform effective retrieval across heterogeneous sources. The irrelevance of retrieved reports affects the factuality of analysis, while insufficient knowledge affects the credibility of clinical decision-making. To bridge the gap, we construct MedAtlas, which includes extensive multimodal report repositories and diverse text corpora. Based on it, we present HeteroRAG, a novel framework that enhances Med-L VLMs through heterogeneous knowledge sources. The framework introduces Modality-specific CLIPs for effective report retrieval and a Multi-corpora Query Generator for dynamically constructing queries for diverse corpora. Incorporating knowledge from such multifaceted sources, Med-L VLM is then trained with Heterogeneous Knowledge Preference Tuning to achieve cross-modality and multi-source knowledge alignment. Extensive experiments across 12 datasets and 3 modalities demonstrate that the proposed HeteroRAG achieves state-of-the-art performance in most medical vision language benchmarks, significantly improving factual accuracy and reliability of Med-L VLMs.

The recent rise in popularity of large language models (LLMs) has prompted considerable concerns about their moral capabilities. Although considerable effort has been dedicated to aligning LLMs with human moral values, existing benchmarks and evaluations remain largely superficial, typically measuring alignment based on final ethical verdicts rather than explicit moral reasoning. In response, this paper aims to advance the investigation of LLMs' moral capabilities by examining their capacity to function as Artificial Moral Assistants (AMAs), systems envisioned in the philosophical literature to support human moral deliberation. We assert that qualifying as an AMA requires more than what state-of-the-art alignment techniques aim to achieve: not only must AMAs be able to discern ethically problematic situations, they should also be able to actively reason about them, navigating between conflicting values outside of those embedded in the alignment phase. Building on existing philosophical literature, we begin by designing a new formal framework of the specific kind of behaviour an AMA should exhibit, individu-ating key qualities such as deductive and abductive moral reasoning. Drawing on this theoretical framework, we develop a benchmark to test these qualities and evaluate popular open LLMs against it. Our results reveal considerable variability across models and highlight persistent shortcomings, particularly regarding abductive moral reasoning. Our work connects theoretical philosophy with practical AI evaluation while also emphasising the need for dedicated strategies to explicitly enhance moral reasoning capabilities in LLMs.