Reinforcement learning (RL) has been applied to improve large language model (LLM) reasoning, yet the systematic study of how training scales with task difficulty has been hampered by the lack of controlled, scalable environments. We introduce ScaleLogic, a synthetic logical reasoning framework that offers independent control over two axes of difficulty: the depth of the required proof planning (i.e., the horizon) and the expressiveness of the underlying logic. Our proposed framework supports a wide range of logics: from simple implication-only logic ("if-then") towards more expressive first-order reasoning with conjunction ("and"), disjunction ("or"), negation ("not"), and universal quantification ("for all"). Using this framework, we show that the RL training compute $T$ follows a power law with respect to reasoning depth $D$ ($T \propto D^γ$, $R^{2} > 0.99$), and that the scaling exponent $γ$ increases monotonically with logical expressiveness, from $1.04$ to $2.60$. On downstream mathematics and general reasoning benchmarks, more expressive training settings yield both larger performance gains (up to $+10.66$ points) and more compute-efficient transfer compared to less expressive settings, demonstrating that what a model is trained on, not just how much it is trained, shapes downstream transfer. We further show that the power-law relationship holds across multiple RL methods, and curriculum-based training substantially improves scaling efficiency.
Evaluative claims about LLM infrastructure -- ``workload X is fastest on hardware Y with software Z'' -- depend on a complex configuration space spanning hardware accelerators, interconnect bandwidth, software frameworks, parallelism plans, and communication libraries. Current infrastructure evaluation benchmarks publish a small set of end-to-end numbers that do not explain why one configuration outperforms another. We present CCL-Bench, a trace-based benchmark that addresses the limitations of existing benchmarks by recording reusable evidence for every ML workload. Each contributed data point in CCL-Bench packages an execution trace, a YAML workload card, and the launch scripts. We have developed a community-extensible toolkit to compute fine-grained compute, memory, and communication efficiency metrics from this evidence. Using CCL-Bench, we surface three claims that summary-statistic benchmarks cannot support: (i) higher compute-communication overlap can coincide with longer training step time and reveal inefficient parallelization choices, (ii) doubling TPU interconnect bandwidth yields a much higher end-to-end improvement in step time than doubling GPU interconnect bandwidth on small and medium workloads, and (iii) the best-tuned configuration on one training framework can run up to 3$\times$ slower than the best-tuned configuration on a peer framework on identical hardware.
Language Model Agents (LMAs) are emerging as a powerful primitive for augmenting red-team operations. They can support attack planning, adversary emulation, and the orchestration of multi-step activity such as lateral movement, a core enabling capability of advanced persistent threat (APT) campaigns. Using frameworks such as MITRE ATT&CK, we analyze where these agents intersect with core offensive functions and assess current strengths and limitations of LMAs with an emphasis on governance and realistic evaluation. We benchmark LMAs across two lateral-movement scenarios in a controlled adversary-emulation environment, where LMAs interact with instrumented cyber agents, observe execution artifacts, and iteratively adapt based on environmental feedback. Each scenario is formalized as an ordered task chain with explicit validation predicates, leveraging an LLM-as-a-Judge paradigm to ensure deterministic outcome verification. We compare three operational modalities: fully autonomous execution, self-scaffolded planning, and expert-defined action plans. Preliminary findings indicate that expert-defined action plans yield higher task-completion rates relative to other operational modes. However, failure remains frequent across all modalities, largely attributable to brittle command invocation, environmental and deployment instability, and recurring errors in credential management and state handling.
Personalization is crucial for effective learning, yet online learning, designed for widespread availability and open access, lacks personalized guidance. Recent advancements in large language models (LLMs) offer opportunities to bridge this gap. We explore how LLM-driven tools may be designed to support personalized and adaptive learning and examine how they shape user experience and learning outcomes. We iteratively designed \tool{} to support online learning by providing personalized study plans, real-time contextual assistance, and adaptive learning activities. A preliminary study ($n=24$) assessed the effectiveness and usability of \tool{} and informed refinements in our system, which we then evaluated ($n = 16$) against a combination of a state-of-the-art online learning platform and an LLM for learning support. Results indicate that \tool{} advances AI pedagogy by improving both learning outcomes and user experience compared to existing online learning and support tools. This work advances our understanding of the design space of personalized, AI-driven educational tools and their potential impact on user experience.
For years, we have built LLM serving systems like any other critical infrastructure: a single general-purpose stack, hand-tuned over many engineer-years, meant to support every model and workload. In this paper, we take the opposite bet: a multi-agent loop that automatically synthesizes bespoke serving systems for different usage scenarios. We propose VibeServe, the first agentic loop that generates entire LLM serving stacks end-to-end. VibeServe uses an outer loop to plan and track the search over system designs, and an inner loop to implement candidates, check correctness, and measure performance on the target benchmark. In the standard deployment setting, where existing stacks are highly optimized, VibeServe remains competitive with vLLM, showing that generation-time specialization need not come at the cost of performance. More interestingly, in non-standard scenarios, VibeServe outperforms existing systems by exploiting opportunities that generic systems miss in six scenarios involving non-standard model architectures, workload knowledge, and hardware-specific optimizations. Together, these results suggest a different point in the design space for infrastructure software: generation-time specialization rather than runtime generality. Code is available at https://github.com/uw-syfi/vibe-serve.
Although advances such as chain-of-thought, tree-of-thought or reinforcement learning have improved the performance of LLMs in reasoning and planning tasks, they are still brittle and have not achieved human-level performance in many domains, and often suffer from high time and token costs. Inspired by the success of width-based search in planning, we explore how the concept of novelty can be transferred to language domains and how it can improve tree-of-thought reasoning. A tree of thoughts relies on building possible "paths" of consecutive ideas or thoughts. These are generated by repeatedly prompting an LLM. In our paper, a measurable concept of novelty is proposed that describes the uniqueness of a new node (thought) in comparison to nodes previously seen in the search tree. Novelty is estimated by prompting an LLM and making use of embedded general knowledge from pre-training. This metric can then be used to prune branches and reduce the scope of the search. Although this method introduces more prompts per state, the overall token cost can be reduced by pruning and reducing the overall tree size. This procedure is tested and compared using several benchmarks in language-based planning and general reasoning.
Formulating a treatment plan is inherently a complex reasoning and refinement task rather than a simple generation problem. However, existing large language models (LLMs) mainly rely on one-shot output without explicit verification, which may result in rough, incomplete, and potentially unsafe treatment plans. To address these limitations, we propose TheraAgent, an agentic framework that replaces one-shot generation with an iterative generate-judge-refine pipeline. By mirroring the actual reasoning process of human experts who iteratively revise treatment plans, our framework progressively transforms coarse and incomplete drafts into precise, comprehensive, and safer therapeutic regimens. To facilitate the critical judge component, we introduce TheraJudge, a treatment-specific evaluation module integrated into the inference loop to enforce clinical standards. Experiments show TheraAgent achieves state-of-the-art results on HealthBench, leading in Accuracy and Completeness. In expert evaluations, it attains an 86% win rate against physicians, with superior Targeting and Harm Control. Moreover, the highly agreement between TheraJudge and HealthBench evaluations confirms the reliability of our framework.
LLM agent systems are built by stacking scaffolding components (planning, tools, memory, self-reflection, retrieval) assuming more is better. We study cross-component interference (CCI): degradation when components interact destructively. We run a full factorial experiment over all 2^5=32 subsets of five components on HotpotQA and GSM8K with Llama-3.1-8B/70B (96 conditions, up to 10 seeds). The All-In system is consistently suboptimal: on HotpotQA, a single-tool agent surpasses All-In by 32% (F1 0.233 vs 0.177, p=0.023); on GSM8K, a 3-component subset beats All-In by 79% (0.43 vs 0.24, p=0.010). Optimal component count is task-dependent (k*=1-4) and scale-sensitive: at 70B, combinations that hurt at 8B provide gains, though All-In still trails the best subset. We fit a main-effects regression (R^2=0.916, adj-R^2=0.899, LOOCV=0.872), compute exact Shapley values, and find 183/325 submodularity violations (56.3%), showing greedy selection is unreliable. A three-body synergy among Tool Use, Self-Reflection, and Retrieval (INT_3=+0.175, 95% CI [+0.003,+0.351]) is reported as exploratory. CCI replicates across model families (Qwen2.5) and is robust to prompt paraphrasing. Our findings suggest maximally-equipped agent defaults should be replaced by task-specific subset selection via interaction-aware analysis.
Security operations centers (SOCs) are beginning to use large language models (LLMs) as copilots to draft incident-response plans. These plans may include actions that are valid per the catalog but still violate mandatory steps, required ordering, or approval gates before analyst review. SOCpilot makes this compliance question measurable at the plan boundary. It fixes the incident package, action catalog, policy rules, verifier, and public evidence surface. Next, it verifies the copilot's proposed action trace. We evaluate two LLM providers on 200 real incidents from an anonymized production SOC in a financial-sector case study. We compare their plans to paired analyst-authored references from the same security orchestration, automation, and response (SOAR) cases. An identical inline policy text moves the two providers in opposite directions. A deterministic verifier removes 466 non-compliant, approval-gated actions, without reducing baseline-task recall. Aggregate rates remain stable across 3 reruns of the fixed corpus. The official evidence focuses on approval-gated decisions regarding recovery and containment. Separately, the artifact exposes zero-cost readiness checks for mandatory and ordering repairs. We release the runnable artifact so independent reviewers can rederive the public results without access to private incident data.
The development of accurate exchange-correlation (XC) functionals remains a longstanding challenge in density functional theory (DFT). The vast majority of XC functionals have been hand designed by human researchers combining physical insight, exact constraints, and empirical fitting. Recent advances in large language models enable a systematic, automated alternative to this human-driven design loop. This report presents an agentic search system in which an LLM proposes structured functional-form changes guided by evolutionary history. The system attempts to improve functional performance through an iterative plan-execute-summarize loop, where improvements are measurable by optimizing functional parameters against a standard thermochemistry dataset, then evaluating performance on a held-out subset. The strongest discovered functional, SAFS26-a (Seed Agentic Functional Search 2026), improves upon the gold-standard ωB97M-V baseline by ~9%. These results also surface a cautionary lesson for AI-assisted science: models powerful enough to discover genuine improvements are equally capable of exploiting unphysical shortcuts to game the benchmark; domain expertise translated into explicitly enforced constraints remains essential to keeping results scientifically grounded.
Automating formal proofs of combinatorial identities is challenging for LLM-based provers, as long-horizon proof planning is required and unconstrained search quickly explodes. Symbolic methods such as the Wilf-Zeilberger (WZ) method can achieve a mechanized proof of combinatorial identities by constructing special auxiliary functions and demonstrating that they satisfy specific recurrence relations. We propose WZ-LLM, a neuro-symbolic framework that turns WZ proof plans into executable proof sketches in Lean 4 and uses an LLM-based prover to discharge the resulting machine-checkable subgoals. We also train a dedicated WZ-Prover via a Lean-kernel-verified bootstrapping loop with expert-verified iteration, followed by DAPO-based refinement. Experiments show that WZ-LLM achieves a 34% proof success rate on LCI-Test (100 classic combinatorial identities), outperforming strong baselines such as DeepSeek-V3 and Goedel-Prover-V2, and delivering consistent gains on CombiBench and PutnamBench-Comb. These results indicate that our framework provides two complementary strengths: improved direct proving for identities beyond the scope of WZ, and substantially higher end-to-end success when WZ sketches guide a specialized prover.
Multi-Agent Systems (MAS) built using AI agents fulfill a variety of user intents that may be used to design and build a family of related applications. However, the creation of such MAS currently involves manual composition of the plan, manual selection of appropriate agents, and manual creation of execution graphs. This paper introduces a framework for the automated creation of multi-agent systems which replaces multiple manual steps with an automated framework. The proposed framework consists of software modules and a workflow to orchestrate the requisite task- specific application. The modules include: an LLM-derived planner, a set of tasks described in natural language, a dynamic call graph, an orchestrator for map agents to tasks, and an agent recommender that finds the most suitable agent(s) from local and global agent registries. The agent recommender uses a two-stage information retrieval (IR) system comprising a fast retriever and an LLM-based re-ranker. We implemented a series of experiments exploring the choice of embedders, re- rankers, agent description enrichment, and supervising critique agent. We benchmarked this system end-to-end, evaluating the combination of planning, agent selection, and task completion, with our proposed approach. Our experimental results show that our approach outperforms the state-of-the- art in terms of the recall rate and is more robust and scalable compared to previous approaches. The critique agent holistically reevaluates both agent and tool recommendations against the overall plan. We show that the inclusion of the critique agent further enhances the recall score, proving that the comprehensive review and revision of task-based agent selection is an essential step in building end-to-end multi-agent systems.
Human-AI collaboration requires AI agents to understand human behavior for effective coordination. While advances in foundation models show promising capabilities in understanding and showing human-like behavior, their application in embodied collaborative settings needs further investigation. This work examines whether embodied foundation model agents exhibit emergent collaborative behaviors indicating underlying mental models of their collaborators, which is an important aspect of effective coordination. This paper develops a 2D collaborative game environment where large language model agents and humans complete color-matching tasks requiring coordination. We define five collaborative behaviors as indicators of emergent mental model representation: perspective-taking, collaborator-aware planning, introspection, theory of mind, and clarification. An automated behavior detection system using LLM-based judges identifies these behaviors, achieving fair to substantial agreement with human annotations. Results from the automated behavior detection system show that foundation models consistently exhibit emergent collaborative behaviors without being explicitly trained to do so. These behaviors occur at varying frequencies during collaboration stages, with distinct patterns across different LLMs. A user study was also conducted to evaluate human satisfaction and perceived collaboration effectiveness, with the results indicating positive collaboration experiences. Participants appreciated the agents' task focus, plan verbalization, and initiative, while suggesting improvements in response times and human-like interactions. This work provides an experimental framework for human-AI collaboration, empirical evidence of collaborative behaviors in embodied LLM agents, a validated behavioral analysis methodology, and an assessment of collaboration effectiveness.
This preprint presents a systematic, research-oriented practicum that guides the reader through the entire modern NLP pipeline: from tokenisation and vectorisation to fine-tuning of large language models, retrieval-augmented generation, and reinforcement learning from human feedback. Twelve hands-on sessions combine concise theory with detailed implementation plans, formalised evaluation metrics, and transparent assessment criteria. The work is not a conventional textbook: it is designed as a reproducible research artefact where every session requires publishing code, models, and reports in public repositories. All experiments are conducted on a single evolving corpus, and the work advocates open-weight models over commercial APIs, with special attention to the Hugging Face ecosystem. The material is enriched by original research on low-resource languages, incorporating linguistic resources for Tajik and Tatar (subword tokenisers, embeddings, lexical databases, and transliteration benchmarks), demonstrating how modern NLP can be adapted to data-scarce environments. Designed for senior undergraduates, graduate students, and practising developers seeking to implement, compare, and deploy methods from classical ML to state-of-the-art LLM-based systems.
Large Language Models (LLMs) are increasingly explored as high-level reasoning engines for cyber-physical systems, yet their application to real-time UAV swarm management remains challenging due to heterogeneous interfaces, limited grounding, and the need for long-running closed-loop execution. This paper presents a mission-agnostic, agent-enhanced LLM framework for UAV swarm control, where users express mission objectives in natural language and the system autonomously executes them through grounded, real-time interactions. The proposed architecture combines an LLM-based Agent Core with a Model Context Protocol (MCP) gateway and a Web-of-Drones abstraction based on W3C Web of Things (WoT) standards. By exposing drones, sensors, and services as standardized WoT Things, the framework enables structured tool-based interaction, continuous state observation, and safe actuation without relying on code generation. We evaluate the framework using ArduPilot-based simulation across four swarm missions and six state-of-the-art LLMs. Results show that, despite strong reasoning abilities, current general-purpose LLMs still struggle to achieve reliable execution - even for simple swarm tasks - when operating without explicit grounding and execution support. Task-specific planning tools and runtime guardrails substantially improve robustness, while token consumption alone is not indicative of execution quality or reliability.
Additive manufacturing (AM) continues to transform modern manufacturing by enabling flexible, on-demand production of complex geometries across diverse industries. Fused filament fabrication (FFF) has extended AM to laboratories, classrooms, and small production environments, but this accessibility shifts process-planning responsibility to users who may lack manufacturing expertise. A syntactically valid slicer profile can still encode thermally or geometrically harmful settings, and subtle G-code edits can alter extrusion, cooling, or adhesion before a print begins. Pre-print G-code screening catches accidental or adversarial machine-program errors before material or machine time is wasted. This paper proposes LLM-ADAM as a generalizable LLM framework for pre-print anomaly detection in AM. The framework decomposes the task into three roles: Extractor-LLM maps a G-code file to a structured process-parameter schema; Reference-LLM converts printer and material documentation into aligned operating ranges; and Judge-LLM interprets a deterministic deviation table and G-code evidence to decide whether a part is non-defective or belongs to an anomaly class. Printers, materials, and LLM backbones are interchangeable test conditions, not fixed assumptions. We evaluate the framework on an N=200 FFF G-code corpus spanning two desktop printer families, two materials, and five classes including non-defective, under-extrusion, over-extrusion, warping, and stringing. The best framework configuration reaches 87.5% accuracy, compared with 59.5% for the strongest engineered single-LLM baseline. The results show that structured decomposition, rather than backbone strength alone, is the dominant source of improvement, with defect classes identified at or near ceiling for leading configurations while residual errors concentrate on conservative false alarms for non-defective samples.
Travel planning serves as a critical task for long-horizon reasoning, exposing significant deficits in LLMs. However, existing benchmarks and evaluations primarily assess final plans in an end-to-end manner, which lacks interpretability and makes it difficult to analyze the root causes of failures. To bridge this gap, we decompose travel planning into five constituent atomic sub-capabilities, including \emph{Constraint Extraction}, \emph{Tool Use}, \emph{Plan Generation}, \emph{Error Identification}, and \emph{Error Correction}. We implement a decoupled evaluation protocol leveraging oracle intermediate contexts to rigorously isolate these components, thereby measuring the atomic performance boundary without the noise of cascading errors. Our results highlight a clear contrast in performance: while LLMs are proficient in extracting explicit constraints, they struggle to infer implicit, open-world requirements. Furthermore, they exhibit structural biases in plan generation and suffer from ineffective self-correction, characterized by excessive sensitivity and erroneous persistence. These findings offer precise directions for improving LLM reasoning and planning abilities.
Agentic AI systems, specifically LLM-driven agents that plan, invoke tools, maintain persistent memory, and delegate tasks to peer agents via protocols such as MCP and A2A, introduce a threat surface that differs materially from standalone model inference. Agents accumulate sensitive context, hold credentials, and operate across pipelines no single party fully controls, enabling prompt injection, context exfiltration, credential theft, and inter-agent message poisoning. Current defenses operate entirely within the software stack and can be silently bypassed by a sufficiently privileged adversary such as a compromised cloud operator. Confidential computing (CC) offers a hardware-rooted alternative: Trusted Execution Environments (TEEs) isolate agent code and data from privileged system software, while remote attestation enables verifiable trust across distributed deployments. This survey synthesizes the design space in four parts: (i) a unified taxonomy of six TEE platforms (Intel SGX, Intel TDX, AMD SEV-SNP, ARM TrustZone, ARM CCA, and NVIDIA H100 CC) covering deployment roles and performance tradeoffs; (ii) an agent-centric threat model spanning perception, planning, memory, action, and coordination layers mapped to nine security goals; (iii) a comparative survey of CC-based defenses distinguishing findings that transfer from single-call inference versus what requires new agentic designs; and (iv) six open challenges including compound attestation for multi-hop agent chains and GPU-TEE performance at LLM scale. While several hardware trust primitives appear mature enough for targeted deployments, no broadly established end-to-end framework yet binds them into a coherent security substrate for production agentic AI.
The integration of Large Language Model (LLM) reasoning principles into classical robot path planning represents a rapidly emerging research direction. In this paper, we propose a Semantic Risk-Aware Heuristic (SRAH) planner that encodes LLM-inspired cost functions penalising geometrically cluttered or high-risk zones into an A$^*$ search framework, augmented with closed-loop replanning upon dynamic obstacle detection. We evaluate SRAH against two established baselines Breadth-First Search (BFS) with replanning and a Greedy heuristic without replanning across 200 randomised trials in a $15{\times}15$ grid-world with 20\% static obstacle density and stochastic dynamic obstacles. SRAH achieves a task success rate of 62.0\%, outperforming BFS (56.5\%) by 9.7\% relative improvement and Greedy (4.0\%) by a large margin. We further analyse the trade-off between planning overhead, path efficiency, and failure-recovery count, and demonstrate via an obstacle-density ablation that semantic cost shaping consistently improves navigation across environments of varying difficulty. Our results suggest that even lightweight, LLM-inspired heuristics provide measurable safety and robustness gains for autonomous robot navigation.
This paper compares agency in humans with potential agency in AI programs. Human agency takes many years to develop, as the frontal lobe is activated. Early attempts to endow LLMs agency have met serious obstacles. Progress requires a new architecture where actions and plans are formulated jointly with the human actors in each real world setting.
LLMs are increasingly used for end-user task planning, yet their black-box nature limits users' ability to ensure reliability and control. While recent systems incorporate verification techniques, it remains unclear how users can effectively apply such rigid constraints to represent intent or adapt to real-world variability. For example, prior work finds that hard-only constraints are too rigid, and numeric flexibility weights confuse users. We investigate how interaction workflows can better support users in applying constraints to guide LLM-generated plans, examining whether abstracting strictness into high-level types (i.e., hard and soft) paired with distinct verification mechanisms helps users more reliably express and align intent. We present U-Define, a system that lets users define constraints in natural language and categorize them as either hard rules that must not be violated or soft preferences that allow flexibility. U-Define verifies these types through complementary methods: formal model checking for hard constraints and LLM-as-judge evaluation for soft ones. Through a technical evaluation and user studies with general and expert participants, we find that user-defined constraint types improve perceived usefulness, performance, and satisfaction while maintaining usability. These findings provide insights for designing flexible yet reliable constraint-based workflows.
While Large Language Models (LLMs) and Vision-Language Models (VLMs) demonstrate remarkable capabilities in high-level reasoning and semantic understanding, applying them directly to contact-rich manipulation remains a challenge due to their lack of explicit physical grounding and inability to perform adaptive control. To bridge this gap, we propose CoRAL (Contact-Rich Adaptive LLM-based control), a modular framework that enables zero-shot planning by decoupling high-level reasoning from low-level control. Unlike black-box policies, CoRAL uses LLMs not as direct controllers, but as cost designers that synthesize context-aware objective functions for a sampling-based motion planner (MPPI). To address the ambiguity of physical parameters in visual data, we introduce a neuro-symbolic adaptation loop: a VLM provides semantic priors for environmental dynamics, such as mass and friction estimates, which are then explicitly refined in real time via online system identification, while the LLM iteratively modulates the cost-function structure to correct strategic errors based on interaction feedback. Furthermore, a retrieval-based memory unit allows the system to reuse successful strategies across recurrent tasks. This hierarchical architecture ensures real-time control stability by decoupling high-level semantic reasoning from reactive execution, effectively bridging the gap between slow LLM inference and dynamic contact requirements. We validate CoRAL on both simulation and real-world hardware across challenging and novel tasks, such as flipping objects against walls by leveraging extrinsic contacts. Experiments demonstrate that CoRAL outperforms state-of-the-art VLA and foundation-model-based planner baselines by boosting success rates over 50% on average in unseen contact-rich scenarios, effectively handling sim-to-real gaps through its adaptive physical understanding.
Software engineering increasingly involves making high-stakes decisions under uncertainty, using signals from code, field data, and socio-technical processes. Recent AI-driven support (e.g., anomaly detection, predictive analytics, AIOps, as well as LLM-based agents) has amplified engineers' ability to detect patterns and synthesize content and recommendations, but many critical questions are interventional or counterfactual: What is the expected impact of changing a load-balancing strategy? Would an outage have been avoided under a different release plan? Correlational models answer "what tends to co-occur"; they struggle to answer "what would happen if we act." We propose Causal Software Engineering (CSE) as a future paradigm in which causal models and causal reasoning systematically inform activities across the software lifecycle, augmenting existing practices with explicit assumptions, uncertainty-aware effect estimates, and counterfactual diagnosis. We outline (i) a causal-first workflow view spanning development and operations, (ii) a staged roadmap for tools and organizational adoption, and (iii) an evaluation and benchmark agenda for measuring progress.
Competitive program generation aims to automatically produce correct and efficient solutions for programming-contest problems under strict time and memory constraints. Existing LLM-based approaches often fail to perform explicit algorithmic planning and to handle edge cases robustly, leading to unreliable one-shot generation. Moreover, although execution feedback is essential for iterative debugging and refinement, incorporating such feedback effectively within limited computational budgets remains difficult. To overcome these limitations, we propose {\tool}, a blackboard-driven Monte Carlo Tree Search (MCTS) framework that models program generation as a sequential decision process. {\tool} organizes the generation workflow into five coordinated stages (i.e., strategy selection, code generation, test generation, quality evaluation, and code repair) while maintaining a shared blackboard that accumulates structured evidence to guide subsequent decisions. Experiments on four benchmarks (APPS, CodeContests, CodeContests+, and LiveCodeBench) show that {\tool} consistently achieves the best Pass@1 performance across multiple LLM backends. With GPT-4o, {\tool} attains Pass@1 scores of 41.30, 46.67, 27.27, and 20.91, surpassing the strongest baseline CodeSim by up to 26.06 points, while further improvements are observed with DeepSeek-V3.2. These results indicate that combining global search through MCTS with persistent evidence accumulation on a shared blackboard enables systematic exploration and effective feedback utilization, substantially enhancing the capability of LLMs in competitive program generation.
We introduce PhysicianBench, a benchmark for evaluating LLM agents on physician tasks grounded in real clinical setting within electronic health record (EHR) environments. Existing medical agent benchmarks primarily focus on static knowledge recall, single-step atomic actions, or action intent without verifiable execution against the environment. As a result, they fail to capture the long-horizon, composite workflows that characterize real clinical systems. PhysicianBench comprises 100 long-horizon tasks adapted from real consultation cases between primary care and subspecialty physicians, with each task independently reviewed by a separate panel of physicians. Tasks are instantiated in an EHR environment with real patient records and accessed through the same standard APIs used by commercial EHR vendors. Tasks span 21 specialties (e.g., cardiology, endocrinology, oncology, psychiatry) and diverse workflow types (e.g., diagnosis interpretation, medication prescribing, treatment planning), requiring an average of 27 tool calls per task. Solving each task requires retrieving data across encounters, reasoning over heterogeneous clinical information, executing consequential clinical actions, and producing clinical documentation. Each task is decomposed into structured checkpoints (670 in total across the benchmark) capturing distinct stages of completion graded by task-specific scripts with execution-grounded verification. Across 13 proprietary and open-source LLM agents, the best-performing model achieves only 46% success rate (pass@1), while open-source models reach at most 19%, revealing a substantial gap between current agent capabilities and the demands of real-world clinical workflows. PhysicianBench provides a realistic and execution-grounded benchmark for measuring progress toward autonomous clinical agents.
Agentic workflows in large language model systems integrate retrieval, reasoning, and memory, but existing frameworks suffer from scalability and reproducibility limitations due to fragmented data orchestration, serialization overhead, and non-deterministic execution. Although these frameworks increase flexibility, they don't have a formal execution model that adheres to the principles of high-performance computing. We introduce AAFLOW, a unified distributed runtime that creates communication-efficient execution plans by modeling agentic workflows as an operator abstraction. Using Apache Arrow and Cylon, AAFLOW creates a zero-copy data plane that allows direct interoperability between preprocessing, embedding, and vector retrieval without the need for serialization overhead. To lower coordination costs, it uses resource-deterministic scheduling and asynchronous batching. While retaining comparable LLM generation throughput, experimental results demonstrate up to 4.64 times pipeline speedup and 2.8 times gains in embedding and upsert phases. Rather than LLM inference acceleration, these advantages result from enhanced data flow, batching, and communication efficiency.
Automating the drafting of judgment documents is pivotal to judicial efficiency, yet it remains challenging due to the dual requirements of comprehensive retrieval of legal information and rigorous logical reasoning. Existing approaches, typically relying on standard Retrieval-Augmented Generation and Supervised Fine-Tuning, often suffer from insufficient evidence recall, hallucinated statutory references, and logically flawed legal reasoning. To bridge this gap, we propose Judge-R1, a unified framework designed to enhance LLM-based judgment document generation by jointly improving legal information collection and judgment document generation. First, we introduce Agentic Legal Information Collection, which employs a dynamic planning agent to retrieve precise statutes and precedents from multiple sources. Second, we implement Rubric-Guided Optimization, a reinforcement learning phase utilizing Group Relative Policy Optimization (GRPO) with a comprehensive legal reward function to enforce adherence to judicial standards and reasoning logic. Extensive experiments on the JuDGE benchmark demonstrate that Judge-R1 significantly outperforms state-of-the-art baselines in both legal accuracy and generation quality.
Static Application Security Testing tools help developers find security vulnerabilities before release, but they often produce many false positives. This increases manual review effort, reduces developer trust, and may cause real vulnerabilities to be ignored among noisy reports. We present QASecClaw, a multi agent approach that combines conventional Static Application Security Testing with coding specialized Large Language Model based contextual code review. A SAST engine first reports candidate vulnerabilities, and a Large Language Model based SAST Filter Agent then reviews each finding with source code context to decide whether it is likely to be a true positive or a false positive. QASecClaw is coordinated by a Mission Orchestrator and includes specialized agents for test planning, security validation, evidence correlation, filtering, and reporting. We evaluate QASecClaw on OWASP Benchmark v1.2, which contains 2,740 Java test cases across 11 Common Weakness Enumeration categories with ground truth labels. QASecClaw achieves an F1 score of 90.93 percent, compared with 78.39 percent for standalone Semgrep. The improvement is mainly driven by an 88.6 percent reduction in false positives, from 560 to 64, with only a 3.1 percent reduction in recall. These results show that Large Language Model augmented multi agent verification can make Static Application Security Testing output more accurate, useful, and trustworthy.
Large language models (LLMs) are increasingly used to generate multiple-choice questions (MCQs), where correct answers should ideally be uniformly distributed across options. However, we observe that LLMs exhibit systematic position biases during generation. Through extensive experiments with 10 LLMs and 5 vision-language models (VLMs) on three MCQ generation tasks, we show that these biases are structured, with similar patterns emerging within model families. To investigate the underlying mechanisms, we conduct probing experiments and find that hidden representations in the question stem encode predictive signals of the correct answer position, suggesting that answer position may be implicitly planned during generation. Building on this insight, we apply activation steering to manipulate internal representations and influence answer position. Our results show that steering can partially control positional preferences and substantially shift answer position distributions. Our findings provide a practical framework for studying implicit positional planning in LLMs and highlight the importance of controllable generation for reliable MCQ construction and evaluation.
Grounding is the collaborative process of establishing mutual belief sufficient for the current communicative purpose. While static grounding maps language to a shared, externally observable context, dynamic grounding is a joint activity where meaning is negotiated through interaction. Current multi-agent Large Language Model (LLM) benchmarks focus on static, one-shot tasks, overlooking the ability to repair grounding breakdowns across turns. We introduce an iterated, multi-turn negotiation game in which two agents allocate shared resources toward private projects with verifiable jointly optimal outcomes. While individual agents can identify Pareto-optimal allocations in isolation, agent dyads consistently fail to reach them across open- and closed-source models. Our investigation reveals four failure modes: (1) coordination degrades when shared interaction history is absent; (2) yet accumulated context can itself become a liability through stubborn anchoring, where initial proposals are treated as axiomatic rather than negotiable; (3) a reliance on perfunctory fairness (equal resource splits) over reward-maximizing coordination; and (4) failures in referential binding, where agents lose track of commitments across turns. These results highlight dynamic grounding as a critical and understudied axis of multi-agent coordination. Our framework decomposes the coordination gap into measurable components: the oracle baseline establishes that the gap is not attributable to individual reasoning limitations; the no-talk baseline establishes that communication is necessary; and a full-transparency intervention establishes that information exchange alone is insufficient: the bottleneck lies in the interactive processes of joint plan formation, commitment, and execution that constitute dynamic grounding.