Large language models increasingly express uncertainty through verbalized confidence, yet these expressions often fail to reflect the model’s sampled behavior. We study verbal confidence alignment as a distributional calibration problem, where the confidence target for a prompt is estimated from repeated model outputs rather than from a single response. However, group rollouts alone are insufficient, since the resulting target must provide a useful training signal. We propose Semantic-Answer Guided Entropy (SAGE), a group-level uncertainty target that constructs an answer-conditioned uncertainty geometry over sampled responses. SAGE preserves categorical, numeric, and symbolic answer distinctions while maintaining a smooth and scale-preserving calibration signal. We further introduce Group-Uncertainty Preference Optimization (GUPO), an uncertainty-channel training framework that supervises verbal confidence expressions rather than the full response. Experiments across factual, mathematical, and multiple-choice reasoning tasks show improved confidence ranking, lower calibration error, and reduced overconfidence.

Large language models (LLMs) have demonstrated remarkable capabilities across diverse tasks, and LLM-based agents further extend these abilities to various practical workflows. While recent progress shows that multi-agent systems (MAS) can outperform single agents by coordinating specialized roles, designing effective MAS remains difficult due to prompt sensitivity and the compounded instability MAS creates. To cope with the challenge, recent efforts in automated prompt design have reduced manual effort. However, multi-agent prompt optimization remains largely unexplored. Challenges like exponentially expanding search space and ambiguous credit assignment together make systematic design intractable without principled methods. Therefore, we introduce Multi-Agent PRompt Optimization (MAPRO), a four-stage framework that first formulates MAS prompt optimization as a Maximum a Posteriori (MAP) inference problem and solves it using a language-guided variant of max-product belief propagation algorithm. To address credit assignment and updates the system iteratively, MAPRO employs a topology-aware refinement mechanism that integrates execution feedback and downstream blames to selectively update agent prompts. Through this process, MAPRO progressively converges to a coordinated set of agent-specific prompt policies. Across benchmarks in various tasks, MAPRO achieves state-of-the-art performance, consistently surpassing manually engineered baselines and recent automated alternatives. Beyond performance, our MAP-based formulation also delivers general guidelines for building more reliable and principled multi-agent systems in the future.

Illicit drug use among teenagers and young adults (TYAs) remains a pressing public health concern, with rising prevalence and long-term impacts on health and well-being. To detect illicit drug use among TYAs, researchers analyze large-scale surveys such as the Youth Risk Behavior Survey (YRBS) and the National Survey on Drug Use and Health (NSDUH), which preserve rich demographic, psychological, and environmental factors related to substance use. However, existing modeling methods treat survey variables independently, overlooking latent and interconnected structures among them. To address this limitation, we propose LAMI (LAtent relation Mining with bi-modal Interpretability), a novel joint graph-language modeling framework for detecting illicit drug use and interpreting behavioral risk factors among TYAs. LAMI represents individual responses as relational graphs, learns latent connections through a specialized graph structure learning layer, and integrates a large language model to generate natural language explanations grounded in both graph structures and survey semantics. Experiments on the YRBS and NSDUH datasets show that LAMI outperforms competitive baselines in predictive accuracy. Interpretability analyses further demonstrate that LAMI reveals meaningful behavioral substructures and psychosocial pathways, such as family dynamics, peer influence, and school-related distress, that align with established risk factors for substance use. Our codebase is available here.

Large Language Models (LLMs) are increasingly integrated into real-world decision-making, including in the domain of public policy. Yet, their ability to comprehend and reason about policy-related content remains underexplored. To fill this gap, we present PolicyBench, the first large-scale bilingual benchmark evaluating policy comprehension, comprising 21K cases across a broad spectrum of policy areas, capturing the diversity and complexity of real-world governance. Following Bloom’s taxonomy, the benchmark assesses three core capabilities: (1) Memorization: factual recall of policy knowledge, (2) Understanding: conceptual and contextual reasoning, and (3) Application: problem-solving in real-life policy scenarios. Building on this benchmark, we further propose PolicyMoE, a domain-specialized Mixture-of-Experts (MoE) model with expert modules aligned to each cognitive level. The proposed models demonstrate stronger performance on application-oriented policy tasks than on memorization or conceptual understanding, and yields the highest accuracy on structured reasoning tasks. Our results reveal key limitations of current LLMs in policy understanding and suggest paths toward more reliable, policy-focused models

Molecular dynamics (MD) is the canonical in-silico method for atomistic molecular science, simulating molecular behavior from first-principles physics. Designing an MD pipeline for a new system requires substantial expert knowledge: running it on even one molecule is expensive, ruling out trial-and-error. We automate this expert pipeline-design process with an LLM agent. Unlike existing MD agents that orchestrate a predefined tool set, we treat pipeline design as open-ended code generation in which the agent’s behavior is reshaped online by verbal reward. Specifically, we build MDForge, an LLM agent whose in-context update rule densifies the sparse reward via a multi-agent debate among physics experts. On three SAMPL host–guest binding free-energy benchmarks (CB[7], OAH, CBClip), MDForge automatically designs MD pipelines competitive with human experts. Deployed on a library of unseen candidate guests, its CB[7] pipeline discovers a novel binder that wet-lab competition NMR confirms is a high-affinity, picomolar CB[7] binder (K_a ≈ 8 × 10¹² M^-1). Our data and code are available at https://github.com/Zehong-Wang/MDForge.

Multi-agent systems built from large language models are deployed widely, yet how much performance is lost when two LLMs must coordinate rather than act alone remains unclear. We formulate the collaboration tax as the team-decentralisation loss of a two-player cooperative game with private information, with two propositions characterising its sign and its equivalence to a max-superadditivity violation. We operationalise this definition on 32 solo-tractable tasks grouped by source of grounding friction and measure it on 11 models from 7 providers. The tax is structured along two no-exception axes: a category ordering across every model and a monotonic decrease with capability. The proximate mechanism is not a reasoning deficit but a four-stage conversational cascade in which agents make ungrounded claims, fail to query the partner, skip integrating both views, and accept the answer without re-derivation. The tax is mechanically predictable from conversation features and partly tractable: a prompt intervention targeting all four stages closes a substantial fraction of the gap, with the dominant bottleneck differing across categories. In heterogeneous pairs the tax is pulled toward the stronger partner rather than the additive midpoint, empirically realising the max-superadditivity violation predicted by our framework. Together these results recast collaboration in LLM systems as a measurable, predictable, and partly tractable cost.

Each LoRA checkpoint compactly stores task-specific updates in low-rank weight matrices, offering an efficient way to adapt large language models to new tasks and domains. In principle, these weights already encode what the adapter does and how well it performs. In this paper, we ask whether this information can be read directly from the weights, without running the base model or accessing training data. A key obstacle is that a single LoRA update can be factorized in infinitely many ways. Without resolving this ambiguity, models trained on the factors may fit the particular factorization rather than the underlying update. To this end, we propose W2T (Weight-to-Token), which maps each LoRA update to a provably canonical form via QR decomposition followed by SVD, so that all equivalent factorizations share the same representation. The resulting components are then tokenized and processed by a Transformer to produce a weight-space embedding. Across language and vision LoRA collections, W2T achieves strong results on attribute classification, performance prediction, and adapter retrieval, demonstrating that LoRA weights reliably indicate model behavior once factorization ambiguity is removed.

As large language models (LLMs) are increasingly deployed in real-world applications, alignment is no longer governed by a single universal notion of safety or helpfulness, but instead by provider- or application-specific model specifications. These specifications are typically long, structured, and frequently updated, yet existing alignment pipelines lack a systematic mechanism to operationalize them as training signals. In this paper, we propose specification-grounded alignment, a new alignment paradigm that treats provider-authored model specifications as the primary alignment target rather than abstract principles or static benchmarks. To instantiate this paradigm, we introduce SpecAlign, a framework that synthesizes alignment data directly from specification documents. SpecAlign combines structured rule annotation, controllable specification instantiation, and multi-agent adversarial data synthesis to generate fine-grained, boundary-aware preference pairs that capture both compliant behaviors and meaningful specification violations. Experiments across multiple model specifications and backbone models demonstrate that training with SpecAlign consistently improves rule compliance while preserving general capabilities and avoiding over-conservative behavior. These results suggest that grounding alignment in explicit model specifications enables rapid, precise, and scalable adaptation of LLM behavior to evolving policy requirements.

Command-line coding agents (e.g., Claude Code, Gemini CLI) can already read and write files and sustain long sessions, yet end-to-end research still fragments across chat tools, IDEs, terminals, and writing environments, and the decisions that make it auditable are rarely preserved. We present **Dr. Claw**, an open-source workspace that wraps existing coding-agent executors in a *controllable* and *auditable* human-in-the-loop workflow rather than introducing another autonomous agent. Persistent state objects, a reusable skill library, and multi-executor coordination link human decisions to AI execution, turning planning, execution, and writing into one traceable, recoverable loop. We demonstrate Dr. Claw through an interactive three-view scenario and a failure-recovery walkthrough, and evaluate it against a bare command-line agent sharing the same backend executor, so the comparison contrasts the whole orchestration layer (task graph, state objects, and skill library) with the agent it wraps. Holding the executor fixed, Dr. Claw scores higher on research completeness while persisting an auditable, recoverable process trail. Demo access: repository https://github.com/OpenLAIR/dr-claw, released under AGPL-3.0 with GPL-3.0 upstream components.

Diet plays a central role in human health, and Nutrition Question Answering (QA) offers a promising path toward personalized dietary guidance and the prevention of diet-related chronic diseases. However, existing methods face two fundamental challenges: the limited reasoning capacity of single-agent systems and the complexity of designing effective multi-agent architectures, as well as contextual overload that hinders accurate decision-making. We introduce Nutritional-Graph Router (NG-Router), a novel framework that formulates nutritional QA as a supervised, knowledge-graph–guided multi-agent collaboration problem. NG-Router integrates agent nodes into heterogeneous knowledge graphs and employs a graph neural network to learn task-aware routing distributions over agents, leveraging soft supervision derived from empirical agent performance. To further address contextual overload, we propose a gradient-based subgraph retrieval mechanism that identifies salient evidence during training, thereby enhancing multi-hop and relational reasoning. Extensive experiments across multiple benchmarks and backbone models demonstrate that NG-Router consistently outperforms both single-agent and ensemble baselines, offering a principled approach to domain-aware multi-agent reasoning for complex nutritional health tasks.

We introduce LongDA, a data analysis benchmark for evaluating LLM-based agents under documentation-intensive analytical workflows. In contrast to existing benchmarks that assume well-specified schemas and inputs, LongDA targets real-world settings in which navigating long documentation and complex data is the primary bottleneck. To this end, we manually curate raw data files, long and heterogeneous documentation, and expert-written publications from 17 publicly available U.S. national surveys, from which we extract 505 analytical queries grounded in real analytical practice. Solving these queries requires agents to first retrieve and integrate key information from multiple unstructured documents before performing multi-step computations and writing executable code, which remains challenging for existing data analysis agents. To support the systematic evaluation under this setting, we develop LongTA, a tool-augmented agent framework that enables document access, retrieval, and code execution, and evaluate a range of proprietary and open-source models. Our experiments reveal substantial performance gaps even among state-of-the-art models, highlighting the challenges researchers should consider before applying LLM agents for decision support in real-world, high-stakes analytical settings.

What enables large language models (LLMs) to effectively model user preferences in sequential recommendation? Our investigation reveals that existing preference-alignment approaches largely rely on binary pairwise comparisons, overlooking two critical factors: preference intensity (the structured strength of affinity or aversion) and temporal context (the extent to which recent interactions better reflect a user’s current intent). Through controlled experiments, we show that leveraging comprehensive feedback with structured preference signals substantially improves recommendation performance, indicating that binary modeling discards essential information. Motivated by these findings, we propose RecPO, a unified preference optimization framework that maps both explicit and implicit feedback into a common preference signal and constructs adaptive reward margins that jointly account for preference intensity and interaction recency. Experiments across five datasets show that RecPO consistently outperforms state-of-the-art baselines while exhibiting behavioral patterns aligned with human decision-making, including favoring immediate satisfaction, maintaining preference coherence, and avoiding dispreferred items. Our results highlight that preference intensity and temporal context are fundamental ingredients for effective LLM-based recommendation. Code: https://github.com/zyouyang/RecPO

Large language models (LLMs) and agent-based frameworks have advanced rapidly, enabling diverse applications. Yet, with the proliferation of models and agentic strategies, practitioners face substantial uncertainty in selecting the best configuration for a downstream task. Prior studies show that different agents and backbones exhibit complementary strengths, and that larger models are not always superior, underscoring the need for adaptive routing mechanisms. Existing approaches to agent routing, however, often emphasize cost efficiency while overlooking the fine-grained contextual and relational structure inherent in QA tasks. In this paper, we propose AgentRouter, a framework that formulates multi-agent QA as a knowledge-graph–guided routing problem supervised by empirical performance signals. Specifically, we convert QA instance into a heterogeneous knowledge graph that jointly encodes queries, contextual entities, and agents, and then train a heterogeneous graph neural network (GNN) to propagate information across node types and produce task-aware routing distributions over agents. By leveraging soft supervision and weighted aggregation of agent outputs, AgentRouter learns principled collaboration schemes that capture the complementary strengths of diverse agents. Extensive experiments demonstrate that our framework consistently outperforms single-agent and ensemble baselines, while generalizing across benchmarks and LLM backbones. These results highlight the effectiveness and robustness of graph-supervised multi-agent routing for question answering. Our code repo is available at https://anonymous.4open.science/r/AgentRouter.

Graphs are ubiquitous structures found in numerous real-world applications, such as drug discovery, recommender systems, and social network analysis. To model graph-structured data, graph neural networks (GNNs) have become a popular tool. However, existing GNN architectures encounter challenges in cross-graph learning where multiple graphs have different feature spaces. To address this, recent approaches introduce text-attributed graphs (TAGs), where each node is associated with a textual description, which can be projected into a unified feature space using textual encoders. While promising, this method relies heavily on the availability of text-attributed graph data, which is difficult to obtain in practice. To bridge this gap, we propose a novel method named Topology-Aware Node description Synthesis (TANS), leveraging large language models (LLMs) to convert existing graphs into text-attributed graphs. The key idea is to integrate topological information into LLMs to explain how graph topology influences node semantics. We evaluate our TANS on text-rich, text-limited, and text-free graphs, demonstrating its applicability. Notably, on text-free graphs, our method significantly outperforms existing approaches that manually design node features, showcasing the potential of LLMs for preprocessing graph-structured data in the absence of textual information. The code and data are available at https://github.com/Zehong-Wang/TANS.

Over the years, online scams have grown dramatically,with nearly 50% of global consumersencountering scam attempts each week.These scams cause not only significant financiallosses to individuals and businesses, butalso lasting psychological trauma, largely dueto scammers’ strategic employment of psychologicaltechniques (PTs) to manipulate victims.Meanwhile, scammers continually evolve theirtactics by leveraging advances in Large LanguageModels (LLMs) to generate diverse scamvariants that easily bypass existing defenses.To address this pressing problem, we introducePsyScam, a benchmark designed to systematicallycapture the PTs employed in real-worldscam reports, and investigate how LLMs canbe utilized to generate variants of scams basedon the PTs and the contexts provided by thesescams. Specifically, we collect a wide range ofscam reports and ground its annotations of employedPTs in well-established cognitive andpsychological theories. We further demonstrateLLMs’ capabilities in generating through twodownstream tasks: scam completion, and scamaugmentation. Experimental results show thatPsyScam presents significant challenges toexisting models in both detecting and generatingscam content based on the PTs used byreal-world scammers. Our code and dataset areavailable.

As the market for illicit drugs remains extremely profitable, major online platforms have become direct-to-consumer intermediaries for illicit drug trafficking participants. These online activities raise significant social concerns that require immediate actions. Existing approaches to combat this challenge are generally impractical due to the scarcity of labeled samples and imbalance of classes in real-world applications. To this end, we propose a novel Large Language Model-empowered Heterogeneous Graph Prompt Learning framework for illicit Drug Trafficking detection, called LLM-HetGDT that leverages LLM to facilitate heterogeneous graph neural networks (HGNNs) to effectively identify minority classes, i.e., drug trafficking participants, in the class-imbalanced scenarios. Specifically, we first pre-train HGNN over a contrastive pretext task to capture the inherent node and structure information over an unlabeled drug trafficking heterogeneous graph (HG). Afterward, to alleviate the class-imbalanced issue, we leverage LLMs to augment the HG by generating high-quality synthetic user nodes in the minority classes. Then, we fine-tune the soft prompts on the augmented HG to capture the important information in the minority classes for the downstream drug trafficking detection task. To comprehensively study online illicit drug trafficking activities, we collect a new HG dataset over Twitter, called Twitter-HetDrug. Extensive experiments on this dataset demonstrate the effectiveness, efficiency, and applicability of our proposed method by comparing it with state-of-the-art baseline methods. Our source code is available at https://github.com/GraphResearcher/LLM-HetGDT.

Diet plays a critical role in human health, yet tailoring dietary reasoning to individual health conditions remains a major challenge. Nutrition Question Answering (QA) has emerged as a popular method for addressing this problem. However, current research faces two critical limitations. On one hand, the absence of datasets involving user-specific medical information severely limits personalization. This challenge is further compounded by the wide variability in individual health needs. On the other hand, while large language models (LLMs), a popular solution for this task, demonstrate strong reasoning abilities, they struggle with the domain-specific complexities of personalized healthy dietary reasoning, and existing benchmarks fail to capture these challenges. To address these gaps, we introduce the Nutritional Graph Question Answering (NGQA) benchmark, the first graph question answering dataset designed for personalized nutritional health reasoning. NGQA leverages data from the National Health and Nutrition Examination Survey (NHANES) and the Food and Nutrient Database for Dietary Studies (FNDDS) to evaluate whether a food is healthy for a specific user, supported by explanations of the key contributing nutrients. The benchmark incorporates three question complexity settings and evaluates reasoning across three downstream tasks. Extensive experiments with LLM backbones and baseline models demonstrate that the NGQA benchmark effectively challenges existing models. In sum, NGQA addresses a critical real-world problem while advancing GraphQA research with a novel domain-specific benchmark. Our codebase and dataset are available here.

A common thread of open-domain question answering (QA) models employs a retriever-reader pipeline that first retrieves a handful of relevant passages from Wikipedia and then peruses the passages to produce an answer. However, even state-of-the-art readers fail to capture the complex relationships between entities appearing in questions and retrieved passages, leading to answers that contradict the facts. In light of this, we propose a novel knowledge graph enhanced passage reader, namely Grape, to improve the reader performance for open-domain QA. Specifically, for each pair of question and retrieved passage, we first construct a localized bipartite graph, attributed to entity embeddings extracted from the intermediate layer of the reader model. Then, a graph neural network learns relational knowledge while fusing graph and contextual representations into the hidden states of the reader model. Experiments on three open-domain QA benchmarks show Grape can improve the state-of-the-art performance by up to 2.2 exact match score with a negligible overhead increase, with the same retriever and retrieved passages. Our code is publicly available at https://github.com/jumxglhf/GRAPE.

We present CoTexT, a pre-trained, transformer-based encoder-decoder model that learns the representative context between natural language (NL) and programming language (PL). Using self-supervision, CoTexT is pre-trained on large programming language corpora to learn a general understanding of language and code. CoTexT supports downstream NL-PL tasks such as code summarizing/documentation, code generation, defect detection, and code debugging. We train CoTexT on different combinations of available PL corpus including both “bimodal” and “unimodal” data. Here, bimodal data is the combination of text and corresponding code snippets, whereas unimodal data is merely code snippets. We first evaluate CoTexT with multi-task learning: we perform Code Summarization on 6 different programming languages and Code Refinement on both small and medium size featured in the CodeXGLUE dataset. We further conduct extensive experiments to investigate CoTexT on other tasks within the CodeXGlue dataset, including Code Generation and Defect Detection. We consistently achieve SOTA results in these tasks, demonstrating the versatility of our models.