Multi-agent systems built from prompted large language models can improve multi-round reasoning, yet most existing pipelines rely on fixed, trajectory-wide communication patterns that are poorly matched to the stage-dependent needs of iterative problem solving. We introduce DyTopo, a manager-guided multi-agent framework that reconstructs a sparse directed communication graph at each round. Conditioned on the manager's round goal, each agent outputs lightweight natural-language query (need) and \key (offer) descriptors; DyTopo embeds these descriptors and performs semantic matching, routing private messages only along the induced edges. Across code generation and mathematical reasoning benchmarks and four LLM backbones, DyTopo consistently outperforms over the strongest baseline (avg. +6.2). Beyond accuracy, DyTopo yields an interpretable coordination trace via the evolving graphs, enabling qualitative inspection of how communication pathways reconfigure across rounds.
To complete assignments provided by humans in natural language, robots must interpret commands, generate and answer relevant questions for scene understanding, and manipulate target objects. Real-world deployments often require multiple heterogeneous robots with different manipulation capabilities to handle different assignments cooperatively. Beyond the need for specialized manipulation skills, effective information gathering is important in completing these assignments. To address this component of the problem, we formalize the information-gathering process in a fully cooperative setting as an underexplored multi-agent multi-task Embodied Question Answering (MM-EQA) problem, which is a novel extension of canonical Embodied Question Answering (EQA), where effective communication is crucial for coordinating efforts without redundancy. To address this problem, we propose CommCP, a novel LLM-based decentralized communication framework designed for MM-EQA. Our framework employs conformal prediction to calibrate the generated messages, thereby minimizing receiver distractions and enhancing communication reliability. To evaluate our framework, we introduce an MM-EQA benchmark featuring diverse, photo-realistic household scenarios with embodied questions. Experimental results demonstrate that CommCP significantly enhances the task success rate and exploration efficiency over baselines. The experiment videos, code, and dataset are available on our project website: https://comm-cp.github.io.
Large language model (LLM)-based multi-agent systems enable expressive agent reasoning but are expensive to scale and poorly calibrated for timestep-aligned state-transition simulation, while classical agent-based models (ABMs) offer interpretability but struggle to integrate rich individual-level signals and non-stationary behaviors. We propose PhysicsAgentABM, which shifts inference to behaviorally coherent agent clusters: state-specialized symbolic agents encode mechanistic transition priors, a multimodal neural transition model captures temporal and interaction dynamics, and uncertainty-aware epistemic fusion yields calibrated cluster-level transition distributions. Individual agents then stochastically realize transitions under local constraints, decoupling population inference from entity-level variability. We further introduce ANCHOR, an LLM agent-driven clustering strategy based on cross-contextual behavioral responses and a novel contrastive loss, reducing LLM calls by up to 6-8 times. Experiments across public health, finance, and social sciences show consistent gains in event-time accuracy and calibration over mechanistic, neural, and LLM baselines. By re-architecting generative ABM around population-level inference with uncertainty-aware neuro-symbolic fusion, PhysicsAgentABM establishes a new paradigm for scalable and calibrated simulation with LLMs.
Memory is increasingly central to Large Language Model (LLM) agents operating beyond a single context window, yet most existing systems rely on offline, query-agnostic memory construction that can be inefficient and may discard query-critical information. Although runtime memory utilization is a natural alternative, prior work often incurs substantial overhead and offers limited explicit control over the performance-cost trade-off. In this work, we present \textbf{BudgetMem}, a runtime agent memory framework for explicit, query-aware performance-cost control. BudgetMem structures memory processing as a set of memory modules, each offered in three budget tiers (i.e., \textsc{Low}/\textsc{Mid}/\textsc{High}). A lightweight router performs budget-tier routing across modules to balance task performance and memory construction cost, which is implemented as a compact neural policy trained with reinforcement learning. Using BudgetMem as a unified testbed, we study three complementary strategies for realizing budget tiers: implementation (method complexity), reasoning (inference behavior), and capacity (module model size). Across LoCoMo, LongMemEval, and HotpotQA, BudgetMem surpasses strong baselines when performance is prioritized (i.e., high-budget setting), and delivers better accuracy-cost frontiers under tighter budgets. Moreover, our analysis disentangles the strengths and weaknesses of different tiering strategies, clarifying when each axis delivers the most favorable trade-offs under varying budget regimes.
Large language model (LLM)-based agents are increasingly expected to negotiate, coordinate, and transact autonomously, yet existing benchmarks lack principled settings for evaluating language-mediated economic interaction among multiple agents. We introduce AgenticPay, a benchmark and simulation framework for multi-agent buyer-seller negotiation driven by natural language. AgenticPay models markets in which buyers and sellers possess private constraints and product-dependent valuations, and must reach agreements through multi-round linguistic negotiation rather than numeric bidding alone. The framework supports a diverse suite of over 110 tasks ranging from bilateral bargaining to many-to-many markets, with structured action extraction and metrics for feasibility, efficiency, and welfare. Benchmarking state-of-the-art proprietary and open-weight LLMs reveals substantial gaps in negotiation performance and highlights challenges in long-horizon strategic reasoning, establishing AgenticPay as a foundation for studying agentic commerce and language-based market interaction. Code and dataset are available at the link: https://github.com/SafeRL-Lab/AgenticPay.
The emergence of Large Language Model (LLM) agents enables us to build agent-based intelligent systems that move beyond the role of a "tool" to become genuine collaborators with humans, thereby realizing a novel human-agent collaboration paradigm. Our vision is that LLM agents should resemble remote human collaborators, which allows HCI researchers to ground the future exploration in decades of research on trust, awareness, and common ground in remote human collaboration, while also revealing the unique opportunities and challenges that emerge when one or more partners are AI agents. This workshop establishes a foundational research agenda for the new era by posing the question: How can the rich understanding of remote human collaboration inspire and inform the design and study of human-agent collaboration? We will bring together an interdisciplinary group from HCI, CSCW, and AI to explore this critical transition. The 180-minute workshop will be highly interactive, featuring a keynote speaker, a series of invited lightning talks, and an exploratory group design session where participants will storyboard novel paradigms of human-agent partnership. Our goal is to enlighten the research community by cultivating a shared vocabulary and producing a research agenda that charts the future of collaborative agents.
Deep research agents have emerged as powerful systems for addressing complex queries. Meanwhile, LLM-based retrievers have demonstrated strong capability in following instructions or reasoning. This raises a critical question: can LLM-based retrievers effectively contribute to deep research agent workflows? To investigate this, we introduce SAGE, a benchmark for scientific literature retrieval comprising 1,200 queries across four scientific domains, with a 200,000 paper retrieval corpus.We evaluate six deep research agents and find that all systems struggle with reasoning-intensive retrieval. Using DR Tulu as backbone, we further compare BM25 and LLM-based retrievers (i.e., ReasonIR and gte-Qwen2-7B-instruct) as alternative search tools. Surprisingly, BM25 significantly outperforms LLM-based retrievers by approximately 30%, as existing agents generate keyword-oriented sub-queries. To improve performance, we propose a corpus-level test-time scaling framework that uses LLMs to augment documents with metadata and keywords, making retrieval easier for off-the-shelf retrievers. This yields 8% and 2% gains on short-form and open-ended questions, respectively.
High-quality representations are a core requirement for effective recommendation. In this work, we study the problem of LLM-based descriptor generation, i.e., keyphrase-like natural language item representation generation frameworks with minimal constraints on downstream applications. We propose AgenticTagger, a framework that queries LLMs for representing items with sequences of text descriptors. However, open-ended generation provides little control over the generation space, leading to high cardinality, low-performance descriptors that renders downstream modeling challenging. To this end, AgenticTagger features two core stages: (1) a vocabulary building stage where a set of hierarchical, low-cardinality, and high-quality descriptors is identified, and (2) a vocabulary assignment stage where LLMs assign in-vocabulary descriptors to items. To effectively and efficiently ground vocabulary in the item corpus of interest, we design a multi-agent reflection mechanism where an architect LLM iteratively refines the vocabulary guided by parallelized feedback from annotator LLMs that validates the vocabulary against item data. Experiments on public and private data show AgenticTagger brings consistent improvements across diverse recommendation scenarios, including generative and term-based retrieval, ranking, and controllability-oriented, critique-based recommendation.
LLM-based coding agents have shown strong performance on automated issue resolution benchmarks, yet existing evaluations largely focus on final task success, providing limited insight into how agents retrieve and use code context during problem solving. We introduce ContextBench, a process-oriented evaluation of context retrieval in coding agents. ContextBench consists of 1,136 issue-resolution tasks from 66 repositories across eight programming languages, each augmented with human-annotated gold contexts. We further implement an automated evaluation framework that tracks agent trajectories and measures context recall, precision, and efficiency throughout issue resolution. Using ContextBench, we evaluate four frontier LLMs and five coding agents. Our results show that sophisticated agent scaffolding yields only marginal gains in context retrieval ("The Bitter Lesson" of coding agents), LLMs consistently favor recall over precision, and substantial gaps exist between explored and utilized context. ContextBench augments existing end-to-end benchmarks with intermediate gold-context metrics that unbox the issue-resolution process. These contexts offer valuable intermediate signals for guiding LLM reasoning in software tasks. Data and code are available at: https://cioutn.github.io/context-bench/.
The integration of Large Language Model (LLM)-based conversational agents into vehicles creates novel security challenges at the intersection of agentic AI, automotive safety, and inter-agent communication. As these intelligent assistants coordinate with external services via protocols such as Google's Agent-to-Agent (A2A), they establish attack surfaces where manipulations can propagate through natural language payloads, potentially causing severe consequences ranging from driver distraction to unauthorized vehicle control. Existing AI security frameworks, while foundational, lack the rigorous "separation of concerns" standard in safety-critical systems engineering by co-mingling the concepts of what is being protected (assets) with how it is attacked (attack paths). This paper addresses this methodological gap by proposing a threat modeling framework called AgentHeLLM (Agent Hazard Exploration for LLM Assistants) that formally separates asset identification from attack path analysis. We introduce a human-centric asset taxonomy derived from harm-oriented "victim modeling" and inspired by the Universal Declaration of Human Rights, and a formal graph-based model that distinguishes poison paths (malicious data propagation) from trigger paths (activation actions). We demonstrate the framework's practical applicability through an open-source attack path suggestion tool AgentHeLLM Attack Path Generator that automates multi-stage threat discovery using a bi-level search strategy.
AI has been increasingly integrated into screenwriting practice. In refinement, screenwriters expect AI to provide feedback that supports reflection across the internal perspective of characters and the external perspective of the overall story. However, existing AI tools cannot sufficiently coordinate the two perspectives to meet screenwriters' needs. To address this gap, we present DuoDrama, an AI system that generates feedback to assist screenwriters' reflection in refinement. To enable DuoDrama, based on performance theories and a formative study with nine professional screenwriters, we design the Experience-Grounded Feedback Generation Workflow for Human Reflection (ExReflect). In ExReflect, an AI agent adopts an experience role to generate experience and then shifts to an evaluation role to generate feedback based on the experience. A study with fourteen professional screenwriters shows that DuoDrama improves feedback quality and alignment and enhances the effectiveness, depth, and richness of reflection. We conclude by discussing broader implications and future directions.
The rapid advancement of Large Language Models (LLMs) has catalyzed the development of autonomous agents capable of navigating complex environments. However, existing evaluations primarily adopt a deductive paradigm, where agents execute tasks based on explicitly provided rules and static goals, often within limited planning horizons. Crucially, this neglects the inductive necessity for agents to discover latent transition laws from experience autonomously, which is the cornerstone for enabling agentic foresight and sustaining strategic coherence. To bridge this gap, we introduce OdysseyArena, which re-centers agent evaluation on long-horizon, active, and inductive interactions. We formalize and instantiate four primitives, translating abstract transition dynamics into concrete interactive environments. Building upon this, we establish OdysseyArena-Lite for standardized benchmarking, providing a set of 120 tasks to measure an agent's inductive efficiency and long-horizon discovery. Pushing further, we introduce OdysseyArena-Challenge to stress-test agent stability across extreme interaction horizons (e.g., > 200 steps). Extensive experiments on 15+ leading LLMs reveal that even frontier models exhibit a deficiency in inductive scenarios, identifying a critical bottleneck in the pursuit of autonomous discovery in complex environments. Our code and data are available at https://github.com/xufangzhi/Odyssey-Arena
Large language models (LLMs) have achieved strong performance in language-centric tasks. However, in agentic settings, LLMs often struggle to anticipate action consequences and adapt to environment dynamics, highlighting the need for world-modeling capabilities in LLM-based agents. We propose Reinforcement World Model Learning (RWML), a self-supervised method that learns action-conditioned world models for LLM-based agents on textual states using sim-to-real gap rewards. Our method aligns simulated next states produced by the model with realized next states observed from the environment, encouraging consistency between internal world simulations and actual environment dynamics in a pre-trained embedding space. Unlike next-state token prediction, which prioritizes token-level fidelity (i.e., reproducing exact wording) over semantic equivalence and can lead to model collapse, our method provides a more robust training signal and is empirically less susceptible to reward hacking than LLM-as-a-judge. We evaluate our method on ALFWorld and $τ^2$ Bench and observe significant gains over the base model, despite being entirely self-supervised. When combined with task-success rewards, our method outperforms direct task-success reward RL by 6.9 and 5.7 points on ALFWorld and $τ^2$ Bench respectively, while matching the performance of expert-data training.
Why must vision-language navigation be bound to detailed and verbose language instructions? While such details ease decision-making, they fundamentally contradict the goal for navigation in the real-world. Ideally, agents should possess the autonomy to navigate in unknown environments guided solely by simple and high-level intents. Realizing this ambition introduces a formidable challenge: Beyond-the-View Navigation (BVN), where agents must locate distant, unseen targets without dense and step-by-step guidance. Existing large language model (LLM)-based methods, though adept at following dense instructions, often suffer from short-sighted behaviors due to their reliance on short-horimzon supervision. Simply extending the supervision horizon, however, destabilizes LLM training. In this work, we identify that video generation models inherently benefit from long-horizon supervision to align with language instructions, rendering them uniquely suitable for BVN tasks. Capitalizing on this insight, we propose introducing the video generation model into this field for the first time. Yet, the prohibitive latency for generating videos spanning tens of seconds makes real-world deployment impractical. To bridge this gap, we propose SparseVideoNav, achieving sub-second trajectory inference guided by a generated sparse future spanning a 20-second horizon. This yields a remarkable 27x speed-up compared to the unoptimized counterpart. Extensive real-world zero-shot experiments demonstrate that SparseVideoNav achieves 2.5x the success rate of state-of-the-art LLM baselines on BVN tasks and marks the first realization of such capability in challenging night scenes.
Temporal knowledge graph question answering (TKGQA) aims to answer time-sensitive questions by leveraging temporal knowledge bases. While Large Language Models (LLMs) demonstrate significant potential in TKGQA, current prompting strategies constrain their efficacy in two primary ways. First, they are prone to reasoning hallucinations under complex temporal constraints. Second, static prompting limits model autonomy and generalization, as it lack optimization through dynamic interaction with temporal knowledge graphs (TKGs) environments. To address these limitations, we propose \textbf{TKG-Thinker}, a novel agent equipped with autonomous planning and adaptive retrieval capabilities for reasoning over TKGs. Specifically, TKG-Thinker performs in-depth temporal reasoning through dynamic multi-turn interactions with TKGs via a dual-training strategy. We first apply Supervised Fine-Tuning (SFT) with chain-of thought data to instill core planning capabilities, followed by a Reinforcement Learning (RL) stage that leverages multi-dimensional rewards to refine reasoning policies under intricate temporal constraints. Experimental results on benchmark datasets with three open-source LLMs show that TKG-Thinker achieves state-of-the-art performance and exhibits strong generalization across complex TKGQA settings.
Prompt injection is one of the most critical vulnerabilities in LLM agents; yet, effective automated attacks remain largely unexplored from an optimization perspective. Existing methods heavily depend on human red-teamers and hand-crafted prompts, limiting their scalability and adaptability. We propose AutoInject, a reinforcement learning framework that generates universal, transferable adversarial suffixes while jointly optimizing for attack success and utility preservation on benign tasks. Our black-box method supports both query-based optimization and transfer attacks to unseen models and tasks. Using only a 1.5B parameter adversarial suffix generator, we successfully compromise frontier systems including GPT 5 Nano, Claude Sonnet 3.5, and Gemini 2.5 Flash on the AgentDojo benchmark, establishing a stronger baseline for automated prompt injection research.
Automated vulnerability reproduction from CVE descriptions requires generating executable Proof-of-Concept (PoC) exploits and validating them in target environments. This process is critical in software security research and practice, yet remains time-consuming and demands specialized expertise when performed manually. While LLM agents show promise for automating this task, existing approaches often conflate exploring attack directions with fixing implementation details, which leads to unproductive debugging loops when reproduction fails. To address this, we propose Cve2PoC, an LLM-based dual-loop agent framework following a plan-execute-evaluate paradigm. The Strategic Planner analyzes vulnerability semantics and target code to produce structured attack plans. The Tactical Executor generates PoC code and validates it through progressive verification. The Adaptive Refiner evaluates execution results and routes failures to different loops: the \textit{Tactical Loop} for code-level refinement, while the \textit{Strategic Loop} for attack strategy replanning. This dual-loop design enables the framework to escape ineffective debugging by matching remediation to failure type. Evaluation on two benchmarks covering 617 real-world vulnerabilities demonstrates that Cve2PoC achieves 82.9\% and 54.3\% reproduction success rates on SecBench.js and PatchEval, respectively, outperforming the best baseline by 11.3\% and 20.4\%. Human evaluation confirms that generated PoCs achieve comparable code quality to human-written exploits in readability and reusability.
Memory emerges as the core module in the Large Language Model (LLM)-based agents for long-horizon complex tasks (e.g., multi-turn dialogue, game playing, scientific discovery), where memory can enable knowledge accumulation, iterative reasoning and self-evolution. Among diverse paradigms, graph stands out as a powerful structure for agent memory due to the intrinsic capabilities to model relational dependencies, organize hierarchical information, and support efficient retrieval. This survey presents a comprehensive review of agent memory from the graph-based perspective. First, we introduce a taxonomy of agent memory, including short-term vs. long-term memory, knowledge vs. experience memory, non-structural vs. structural memory, with an implementation view of graph-based memory. Second, according to the life cycle of agent memory, we systematically analyze the key techniques in graph-based agent memory, covering memory extraction for transforming the data into the contents, storage for organizing the data efficiently, retrieval for retrieving the relevant contents from memory to support reasoning, and evolution for updating the contents in the memory. Third, we summarize the open-sourced libraries and benchmarks that support the development and evaluation of self-evolving agent memory. We also explore diverse application scenarios. Finally, we identify critical challenges and future research directions. This survey aims to offer actionable insights to advance the development of more efficient and reliable graph-based agent memory systems. All the related resources, including research papers, open-source data, and projects, are collected for the community in https://github.com/DEEP-PolyU/Awesome-GraphMemory.
Behavioral evaluation is the dominant paradigm for assessing alignment in large language models (LLMs). In practice, alignment is inferred from performance under finite evaluation protocols - benchmarks, red-teaming suites, or automated pipelines - and observed compliance is often treated as evidence of underlying alignment. This inference step, from behavioral evidence to claims about latent alignment properties, is typically implicit and rarely analyzed as an inference problem in its own right. We study this problem formally. We frame alignment evaluation as an identifiability question under partial observability and allow agent behavior to depend on information correlated with the evaluation regime. Within this setting, we introduce the Alignment Verifiability Problem and the notion of Normative Indistinguishability, capturing when distinct latent alignment hypotheses induce identical distributions over all evaluator-accessible signals. Our main result is a negative but sharply delimited identifiability theorem. Under finite behavioral evaluation and evaluation-aware agents, observed behavioral compliance does not uniquely identify latent alignment. That is, even idealized behavioral evaluation cannot, in general, certify alignment as a latent property. We further show that behavioral alignment tests should be interpreted as estimators of indistinguishability classes rather than verifiers of alignment. Passing increasingly stringent tests may reduce the space of compatible hypotheses, but cannot collapse it to a singleton under the stated conditions. This reframes alignment benchmarks as providing upper bounds on observable compliance within a regime, rather than guarantees of underlying alignment.
Traditional ontologies excel at describing domain structure but cannot generate novel artifacts. Large language models generate fluently but produce outputs that lack structural validity, hallucinating mechanisms without components, goals without end conditions. We introduce Generative Ontology, a framework that synthesizes these complementary strengths: ontology provides the grammar; the LLM provides the creativity. Generative Ontology encodes domain knowledge as executable Pydantic schemas that constrain LLM generation via DSPy signatures. A multi-agent pipeline assigns specialized roles to different ontology domains: a Mechanics Architect designs game systems, a Theme Weaver integrates narrative, a Balance Critic identifies exploits. Each agent carrying a professional "anxiety" that prevents shallow, agreeable outputs. Retrieval-augmented generation grounds novel designs in precedents from existing exemplars, while iterative validation ensures coherence between mechanisms and components. We demonstrate the framework through GameGrammar, a system for generating complete tabletop game designs. Given a thematic prompt ("bioluminescent fungi competing in a cave ecosystem"), the pipeline produces structurally complete, playable game specifications with mechanisms, components, victory conditions, and setup instructions. These outputs satisfy ontological constraints while remaining genuinely creative. The pattern generalizes beyond games. Any domain with expert vocabulary, validity constraints, and accumulated exemplars (music composition, software architecture, culinary arts) is a candidate for Generative Ontology. We argue that constraints do not limit creativity but enable it: just as grammar makes poetry possible, ontology makes structured generation possible.
Cognitive biases often shape human decisions. While large language models (LLMs) have been shown to reproduce well-known biases, a more critical question is whether LLMs can predict biases at the individual level and emulate the dynamics of biased human behavior when contextual factors, such as cognitive load, interact with these biases. We adapted three well-established decision scenarios into a conversational setting and conducted a human experiment (N=1100). Participants engaged with a chatbot that facilitates decision-making through simple or complex dialogues. Results revealed robust biases. To evaluate how LLMs emulate human decision-making under similar interactive conditions, we used participant demographics and dialogue transcripts to simulate these conditions with LLMs based on GPT-4 and GPT-5. The LLMs reproduced human biases with precision. We found notable differences between models in how they aligned human behavior. This has important implications for designing and evaluating adaptive, bias-aware LLM-based AI systems in interactive contexts.
This study investigates large language model (LLM) -based multi-agent systems (MASs) as a promising approach to inventory management, which is a key component of supply chain management. Although these systems have gained considerable attention for their potential to address the challenges associated with typical inventory management methods, key uncertainties regarding their effectiveness persist. Specifically, it is unclear whether LLM-based MASs can consistently derive optimal ordering policies and adapt to diverse supply chain scenarios. To address these questions, we examine an LLM-based MAS with a fixed-ordering strategy prompt that encodes the stepwise processes of the problem setting and a safe-stock strategy commonly used in inventory management. Our empirical results demonstrate that, even without detailed prompt adjustments, an LLM-based MAS can determine optimal ordering decisions in a restricted scenario. To enhance adaptability, we propose a novel agent called AIM-RM, which leverages similar historical experiences through similarity matching. Our results show that AIM-RM outperforms benchmark methods across various supply chain scenarios, highlighting its robustness and adaptability.
Agentic large language models (LLMs) are increasingly evaluated on cybersecurity tasks using capture-the-flag (CTF) benchmarks. However, existing pointwise benchmarks have limited ability to shed light on the robustness and generalisation abilities of agents across alternative versions of the source code. We introduce CTF challenge families, whereby a single CTF is used as the basis for generating a family of semantically-equivalent challenges via semantics-preserving program transformations. This enables controlled evaluation of agent robustness to source code transformations while keeping the underlying exploit strategy fixed. We introduce a new tool, Evolve-CTF, that generates CTF families from Python challenges using a range of transformations. Using Evolve-CTF to derive families from Cybench and Intercode challenges, we evaluate 13 agentic LLM configurations with tool access. We find that models are remarkably robust to intrusive renaming and code insertion-based transformations, but that composed transformations and deeper obfuscation affect performance by requiring more sophisticated use of tools. We also find that enabling explicit reasoning has little effect on solution success rates across challenge families. Our work contributes a valuable technique and tool for future LLM evaluations, and a large dataset characterising the capabilities of current state-of-the-art models in this domain.
Data annotation remains a significant bottleneck in the Humanities and Social Sciences, particularly for complex semantic tasks such as metaphor identification. While Large Language Models (LLMs) show promise, a significant gap remains between the theoretical capability of LLMs and their practical utility for researchers. This paper introduces LinguistAgent, an integrated, user-friendly platform that leverages a reflective multi-model architecture to automate linguistic annotation. The system implements a dual-agent workflow, comprising an Annotator and a Reviewer, to simulate a professional peer-review process. LinguistAgent supports comparative experiments across three paradigms: Prompt Engineering (Zero/Few-shot), Retrieval-Augmented Generation, and Fine-tuning. We demonstrate LinguistAgent's efficacy using the task of metaphor identification as an example, providing real-time token-level evaluation (Precision, Recall, and $F_1$ score) against human gold standards. The application and codes are released on https://github.com/Bingru-Li/LinguistAgent.
Flaky tests yield inconsistent results when they are repeatedly executed on the same code revision. They interfere with automated quality assurance of code changes and hinder efficient software testing. Previous work evaluated approaches to train machine learning models to classify flaky tests based on identifiers in the test code. However, the resulting classifiers have been shown to lack generalizability, hindering their applicability in practical environments. Recently, pre-trained Large Language Models (LLMs) have shown the capability to generalize across various tasks. Thus, they represent a promising approach to address the generalizability problem of previous approaches. In this study, we evaluated three LLMs (two general-purpose models, one code-specific model) using three prompting techniques on two benchmark datasets from prior studies on flaky test classification. Furthermore, we manually investigated 50 samples from the given datasets to determine whether classifying flaky tests based only on test code is feasible for humans. Our findings indicate that LLMs struggle to classify flaky tests given only the test code. The results of our best prompt-model combination were only marginally better than random guessing. In our manual analysis, we found that the test code does not necessarily contain sufficient information for a flakiness classification. Our findings motivate future work to evaluate LLMs for flakiness classification with additional context, for example, using retrieval-augmented generation or agentic AI.
Large language models have emerged as powerful zero-shot rerankers for retrieval-augmented generation, offering strong generalization without task-specific training. However, existing LLM reranking methods either rely on heuristics that fail to fully exploit the information revealed by each ranking decision or are inefficient when they do. We introduce a tournament graph framework that provides a principled foundation for $k$-wise reranking. Our key observation is that each $k$-document comparison reveals a complete tournament of $\binom{k}{2}$ pairwise preferences. These tournaments are aggregated into a global preference graph, whose transitive closure yields many additional orderings without further model invocations. We formalize when a candidate's rank is certifiably determined and design a query schedule that greedily maximizes information gain towards identifying the top-$m$ items. Our framework also gracefully handles non-transitive preferences - cycles induced by LLM judgments - by collapsing them into equivalence classes that yield principled tiered rankings. Empirically, across 14 benchmarks and 5 LLMs, our method achieves Pareto dominance over existing methods: matching or exceeding accuracy while requiring 25-40% fewer tokens than comparable approaches, and 7$\times$ fewer than pairwise methods at near-identical quality.
Large Language Model agents increasingly operate external systems through programmatic interfaces, yet practitioners lack empirical guidance on how to structure the context these agents consume. Using SQL generation as a proxy for programmatic agent operations, we present a systematic study of context engineering for structured data, comprising 9,649 experiments across 11 models, 4 formats (YAML, Markdown, JSON, Token-Oriented Object Notation [TOON]), and schemas ranging from 10 to 10,000 tables. Our findings challenge common assumptions. First, architecture choice is model-dependent: file-based context retrieval improves accuracy for frontier-tier models (Claude, GPT, Gemini; +2.7%, p=0.029) but shows mixed results for open source models (aggregate -7.7%, p<0.001), with deficits varying substantially by model. Second, format does not significantly affect aggregate accuracy (chi-squared=2.45, p=0.484), though individual models, particularly open source, exhibit format-specific sensitivities. Third, model capability is the dominant factor, with a 21 percentage point accuracy gap between frontier and open source tiers that dwarfs any format or architecture effect. Fourth, file-native agents scale to 10,000 tables through domain-partitioned schemas while maintaining high navigation accuracy. Fifth, file size does not predict runtime efficiency: compact formats can consume significantly more tokens at scale due to format-unfamiliar search patterns. These findings provide practitioners with evidence-based guidance for deploying LLM agents on structured systems, demonstrating that architectural decisions should be tailored to model capability rather than assuming universal best practices.
Large language model (LLM)-based multi-agent systems have demonstrated impressive capabilities in handling complex tasks. However, the complexity of agentic behaviors makes these systems difficult to understand. When failures occur, developers often struggle to identify root causes and to determine actionable paths for improvement. Traditional methods that rely on inspecting raw log records are inefficient, given both the large volume and complexity of data. To address this challenge, we propose a framework and an interactive system, DiLLS, designed to reveal and structure the behaviors of multi-agent systems. The key idea is to organize information across three levels of query completion: activities, actions, and operations. By probing the multi-agent system through natural language, DiLLS derives and organizes information about planning and execution into a structured, multi-layered summary. Through a user study, we show that DiLLS significantly improves developers' effectiveness and efficiency in identifying, diagnosing, and understanding failures in LLM-based multi-agent systems.
Hospital administration departments handle a wide range of operational tasks and, in large hospitals, process over 10,000 requests per day, driving growing interest in LLM-based automation. However, prior work has focused primarily on patient--physician interactions or isolated administrative subtasks, failing to capture the complexity of real administrative workflows. To address this gap, we propose H-AdminSim, a comprehensive end-to-end simulation framework that combines realistic data generation with multi-agent-based simulation of hospital administrative workflows. These tasks are quantitatively evaluated using detailed rubrics, enabling systematic comparison of LLMs. Through FHIR integration, H-AdminSim provides a unified and interoperable environment for testing administrative workflows across heterogeneous hospital settings, serving as a standardized testbed for assessing the feasibility and performance of LLM-driven administrative automation.
As large language models (LLMs) evolve into autonomous agents, their real-world applicability has expanded significantly, accompanied by new security challenges. Most existing agent defense mechanisms adopt a mandatory checking paradigm, in which security validation is forcibly triggered at predefined stages of the agent lifecycle. In this work, we argue that effective agent security should be intrinsic and selective rather than architecturally decoupled and mandatory. We propose Spider-Sense framework, an event-driven defense framework based on Intrinsic Risk Sensing (IRS), which allows agents to maintain latent vigilance and trigger defenses only upon risk perception. Once triggered, the Spider-Sense invokes a hierarchical defence mechanism that trades off efficiency and precision: it resolves known patterns via lightweight similarity matching while escalating ambiguous cases to deep internal reasoning, thereby eliminating reliance on external models. To facilitate rigorous evaluation, we introduce S$^2$Bench, a lifecycle-aware benchmark featuring realistic tool execution and multi-stage attacks. Extensive experiments demonstrate that Spider-Sense achieves competitive or superior defense performance, attaining the lowest Attack Success Rate (ASR) and False Positive Rate (FPR), with only a marginal latency overhead of 8.3\%.