Although existing candidate trajectory evaluators have substantially improved end-to-end planning accuracy, their underlying mechanism remains suboptimal. Model-based evaluators typically make scoring future-aware by explicitly predicting candidate-conditioned future states over the planning horizon, which we term rollout. However, rollout incurs high inference cost and accumulates prediction errors. We argue that planning does not require reconstructing a full future scene for each candidate, but only future-relevant information for reliable comparison. We therefore propose RFWorld, a Rollout-Free World model for end-to-end trajectory evaluation. RFWorld constructs a shared scene memory and shapes it into a future-readable representation via lightweight training-time temporal readout with semantic supervision, keeping training overhead low. At inference time, candidate trajectories directly query the shared memory via sparse spatial readout for scoring, reducing computation by roughly the planning horizon times the number of candidate trajectories, while avoiding recursive rollout errors. Experiments on NAVSIM and Bench2Drive show that RFWorld learns a more planning-aligned world representation and achieves state-of-the-art performance.
Existing RAW-based detectors typically introduce learnable ISP modules to convert RAW inputs into sRGB-like formats to fit off-the-shelf detectors. However, this RAW-to-sRGB paradigm decouples representation learning from the downstream detection task. Crucially, it forces the network to inherit human-centric visual priors while discarding intrinsic RAW advantages, such as linear sensor responses and high-bit-depth, minimally processed scene cues. In this paper, we propose \textbf{ILLRAW}, an \textbf{ILL}umination-aware end-to-end \textbf{RAW} object detection framework. Instead of reconstructing visually plausible sRGB images, ILLRAW directly learns detection-oriented RAW features through a lightweight feature adapter inserted at the input stage of existing detectors. ILLRAW estimates an illumination mask and introduces channel-sensitive embeddings to model Bayer-channel responses. These features are fused by an illumination attention fusion module. Experiments show clear improvements on the diverse AODRaw benchmark and competitive performance on LOD and PascalRAW, with particularly strong results under dark exposure.
Recent advances in video generation have enabled generative video compression to reconstruct details lost during coding using pretrained spatiotemporal priors. However, a substantial gap remains between codec-induced degradation and the perturbation trajectories learned by flow-matching models. In this paper, we identify a compression-dependent directional mismatch that degradation-strength-based timestep calibration alone cannot resolve. Compressed latents remain strongly aligned across codec operating points, while their reconstruction residuals vary substantially in magnitude and direction. To bridge this gap, we propose RIFT, a Residual-Informed Flow Transfer framework for one-step generative video compression. At its core, RIFT introduces a Residual Geometry Inducer (RGI) that converts the pretrained flow prediction into a content-adaptive restoration update using decoder-visible codec cues. RGI integrates flow-timestep conditioning with a learned full-band residual-strength estimate and content-conditioned direction and gain modulation, aligning the orientation and magnitude of the update with the compression-dependent residual geometry. Extensive experiments on three public video compression benchmarks show that RIFT outperforms competing generative video compression methods. The code is publicly available on RIFT.
Zizhe Wang, Jia Deng, Xin Lu, Zehao Jia, Yi Huang, Yao Zhu, Jianing Li
Artificial intelligence for science (AI4S) is extending from interpretation of existing data to the design and execution of experiments. Space offers distinctive scientific opportunities while restricting physical support. Microgravity and extraterrestrial environments expose phenomena that terrestrial experiments cannot fully reproduce; finite samples, constrained resources, and limited opportunities for intervention complicate sustained investigation. This critical review examines orbital laboratories and planetary in situ science across biology, materials, fluid and thermal physics, and adaptive exploration. We distinguish learning from space-derived data, selecting scientific actions, executing them, and preserving experimental validity. Terrestrial self-driving laboratories provide complementary evidence on optimization, dedicated automation, robotic manipulation, and the support needed for unattended operation. The synthesis connects scientific gains already demonstrated by bounded AI systems to gaps in observation, executable actions, and continuity of valid evidence. Space robots can expand the attainable scope of autonomous science when necessary interventions exceed validated instrument capabilities and timely human support is unavailable. Their value must be established against dedicated mechanisms, redundancy, and human-supported operation. We develop a framework connecting sample history, scientific decisions, physical execution, and claim-specific validity. A worked orbital design, a planetary extension, and matched evaluation schemes separate the contributions of sensing, decision policies, and physical actions. The resulting agenda measures autonomy through valid scientific return, task coverage, intervention costs, and evidence integrity across complete campaigns.
Junnan Nie, Zitian Fan, Jiachen Zhang, Junyi Lao, Likui Zhang, Chenghao Liu, Xiaoyang Lu, Yunbo Wang 等 17 位作者
Learning task progress from robot demonstrations provides a practical way to construct dense shaping rewards for robotic reinforcement learning. A common approach uses normalized demonstration time as progress labels, implicitly assuming monotonically increasing task progress. During policy exploration, however, actual task progress can stall or even regress. Such models may mistake regressions for progress, assigning positive shaping rewards to actions that cause progress loss. Real suboptimal and failed executions offer corrective experience, but assigning absolute progress values to their states is difficult. Our key observation is that stalls, regressions, and recoveries establish local relative progress relations even when absolute progress is unknown. We introduce Event2Reward, which uses normalized temporal labels to anchor the progress scale and these within-trajectory relations to constrain local changes, combining absolute and relative progress supervision. Its progress predictions serve as potentials for reward shaping. We curate an event-annotated dataset from approximately 100K raw real-robot execution records, covering successful, suboptimal, and failed executions, without requiring framewise absolute progress labels. Under the same training budgets, Event2Reward improves average policy success by 18.16 percentage points over a baseline using only sparse task rewards across 40 tasks in four LIBERO suites. On real robots, average success after 30 minutes of wall-clock training reaches 60.0% on three precise insertion tasks and 90.0% on three object manipulation tasks, exceeding the sparse-reward baseline under the same budget by 36.7 and 23.3 percentage points, respectively. Ablation studies further show that absolute progress supervision and relative progress relations are complementary, with joint supervision yielding better downstream policy learning.
Visual foundation models learn rich pathological morphology but lack explicit concept-level supervision. Vision--language pretraining can provide this supervision but requires costly image--text pairs. Here, we introduce GloSAGE, a semantic self-distillation framework that uses lesion profiles predicted by a pretrained visual teacher to construct structured pathology descriptions. Reliability-guided selection identifies image--text anchors, while Knowledge-Structured Pathology Text Synthesis and Evidence-Guided Negative-pair Construction account for partially observed lesions. Image--text alignment trains projection layers while retaining the frozen visual encoder; an intermediate Semantic Transfer Layer supplies features for downstream tasks. Developed on GloMillion and evaluated across seven glomerulus- and case-level cohorts, GloSAGE supported zero-shot lesion recognition, image--text retrieval, hard-negative discrimination and concept-intervention analyses. It outperformed its source visual model in most lesion-recognition tasks and improved lesion grading, while supporting competitive case-level disease classification, injury grading and baseline renal function estimation. Ablations supported the combined contributions of knowledge-structured text and evidence-guided negatives, and the intermediate transfer layer improved mean lesion-recognition ROC-AUC relative to the complete projection output. A frozen language model conditioned on GloSAGE features generated glomerular descriptions. These findings suggest that teacher-derived lesion profiles can provide annotation-efficient semantic supervision for extending a glomerular visual foundation model.
Rui Tian, Yi Fang, Yejin Hua, Du Wang, Yuzhe Wang, Minxuan Shan, Zirui Ren, Xinyu Lin 等 11 位作者
Recommender Systems Engineer Agents (RSE-Agents), which use LLM-based agents to autonomously improve recommendation algorithms, are emerging as a promising paradigm. However, their capabilities and limitations remain poorly understood due to the lack of a unified, standardized evaluation framework and systematic trajectory-level analysis. To address this gap, we introduce AutoRSE, an evaluation framework spanning six representative industrial-grade recommendation tasks, and conduct a large-scale empirical study of three frontier LLMs across three widely used agent harnesses. Our results show that RSE-Agents consistently enhance the performance of backbones on five of the six tasks, with an average relative improvement of 16.1%, indicating that they can already autonomously improve recommendation performance. However, trajectory-level analysis also reveals that current RSE-Agents function primarily as engineering optimizers rather than algorithmic improvers: their gains are largely driven by hyperparameter tuning, training strategies, and model ensembling, while genuine algorithmic innovations are rare. Since the largest gains in recommender systems often stem from algorithmic advances, this behavior underscores a fundamental limitation of current RSE-Agents. To probe this limitation, we perform idea-injection experiments and show that agents equipped with innovative ideas from frontier research or human experts achieve substantially larger improvements than those relying on self-generated ideas. These results suggest that idea generation is a central bottleneck of current RSE-Agents, highlights a key direction for the future recommendation system research. Our implementation is publicly available on GitHub: https://github.com/Ray-Tiann/AutoRSE.
Hannah P. Menke, Ahmed H. Elsheikh, Lingli Wei, Nanzhe Wang, Andreas Busch
Geological CO2 storage, geothermal energy and underground hydrogen storage depend on irreversible or high-cost decisions made with sparse data and incomplete physical understanding. Machine learning (ML) increasingly informs these decisions, yet methods are usually judged on predictive accuracy rather than on whether their outputs can be trusted, traced and acted upon. We argue that subsurface ML should be designed as decision infrastructure with shared systems that carry calibrated uncertainty and provenance from observation to operating decision. We propose a decision-readiness checklist, an audit-trail workflow for monitoring updates, and benchmark designs that score models by the decisions they support. We identify validation needs shared across the three technologies, and the data-access and governance steps needed for ML to support regulatory and investment decisions.
Zehao Wang, Hanzhu Chen, Runquan Gui, Zhe Wang, Lin Yang
Reinforcement Learning with Verifiable Rewards (RLVR) derives its learning signal from binary rewards via on-policy exploration and becomes a powerful paradigm for improving the reasoning of large language models. Yet it faces a learning cliff on problems beyond model's current ability: when every rollout fails, the samples reveal nothing about a correct solution, and group-relative estimators (e.g., GRPO, RLOO) collapse the update to zero advantage. Expert trajectories offer a remedy, but they are incompatible with the current policy and often lie in the student's low-probability regions, so the surviving updates demand large parameter shifts and decay into surface-level imitation rather than transferable reasoning. We argue that a useful learning signal on hard problems must be both verifiably correct and sufficiently compatible with the current policy. We propose Policy-Compatible Reinforcement Learning (PICO), which performs distribution-constrained joint decoding: expert data guides decoding at the token level while preserving the student's relative token probabilities, and a dynamic top-k mechanism confines the sampling distribution under a finite divergence budget, yielding trajectories that are verified and policy-compatible by construction. These trajectories are optimized jointly with on-policy rollouts under a mixed-behavior objective that mitigates their distribution mismatch via importance sampling. Experiments across reasoning diverse datasets and agentic benchmarks demonstrate that PICO consistently outperforms standard RLVR and prior off-policy baselines.
Humanoid dexterous hands are a core component of embodied intelligence, enabling robots to translate high-level intent into rich and adaptive physical interaction. However, existing designs remain constrained by a persistent trade-off among anthropomorphic kinematic complexity, mechanical robustness, load capacity and control simplicity, limiting dexterous manipulation in unstructured environments. Here we present Torch-Hand, a fully articulated humanoid dexterous hand that replicates the joint structure and motion hierarchy of the human hand with 21 degrees of freedom (DOF), comprising a highly articulated five-DOF thumb and four DOF in each of the remaining fingers. The differentiated thumb integrates multi-axis CMC motion to provide broad opposition and flexible contact orientation relative to the four fingers. Unlike tendon-driven or hybrid actuation approaches, Torch-Hand adopts an entirely linkage-based transmission architecture. By integrating screw-linkage actuation, the hand achieves high grasping force and substantial payload capacity while maintaining a compact anthropomorphic form. Its actuation architecture is optimized for control tractability, with all actuators integrated within the palm and arranged in a serial-robot-like configuration that establishes a direct correspondence between motor motion and joint kinematics. This architecture mechanically decouples functional DOF to reduce unintended inter-joint motion propagation, while allowing independently addressable DOF to be actively coordinated into coupled multi-joint motion patterns during complex manipulation. Comprehensive experiments assess reachable workspace, coordinated fingertip trajectories, thumb opposition, 33-grasp taxonomy coverage, grasping force, payload capacity and task-level manipulation. These results establish Torch-Hand as a robust hardware foundation for dexterous embodied manipulation.
Command-and-control, data transfer, and censorship circumvention can hide in traffic to widely used services. DNS-bound access control grants permits based on resolved IP addresses, while shared edges route requests by tenant identity, creating a validation gap. An attacker can resolve an allowed anchor, obtain a permit for its shared edge address, and access a hidden tenant by presenting the tenant's name as both SNI and HTTP Host. To investigate this gap, we present the Domain Decoupling Attack (DDA), which exploits this gap in CDN and non-CDN hosting while preserving SNI--Host consistency. Across 1,069,048 domains, we conduct 18,025,068 successful probes. Overall, 95.8% of domains accept an external SNI, and 84.6% return content matching the named tenant. Exposure reaches 99.26% across 506,763 CDN domains and 92.75% across 562,285 non-CDN domains. The co-location graph places 666,413 exposed domains in cross-tenant components; 97.7% of 54,069 cross-tenant non-CDN addresses have a confirmed path to another tenant. None of the tested defenses bind tenant identity to the DNS-approved name. Cleartext inline IPS can block access based on SNI, but Encrypted Client Hello (ECH) hides it. The Windows Zero Trust DNS (ZTDNS) permit remains reusable in both cases, enabling the concealment of command-and-control and data-exfiltration traffic and facilitating censorship circumvention. The risk conceals command-and-control and data-exfiltration traffic and supports censorship.
Hongcheng Gao, Zelin Zheng, Alex Zhou, Jay Zhu, Shijia Ge, Yazhe Wang, Xiping Li, Jianshu Zeng 等 21 位作者
Embodied agents are expected to accomplish real-world goals by deciding for themselves how and when to interact with their surroundings. For many such goals, changing the physical state of the world is the means rather than the end: the agent changes the state of a system through interaction, observes the feedback, and explores until it either achieves the goal itself (e.g., discovering physical laws) or learns what the goal requires; it acts to know. Manipulation benchmarks stop at acting to change, with the target state given; scientific discovery benchmarks evaluate knowing but replace physical experimentation with symbolic actions or numerical queries. We introduce EmbodiedLab, a benchmark in which a robot must act to know. Quantitative science experiments make this setting controllable: the parameters of the system are resampled for every instance, so the result cannot be known in advance, yet it can be scored exactly. Across 30 experiments in five disciplines on 14 apparatus platforms, the agent receives an objective and the apparatus, chooses its own procedure, and operates a physics-based simulated laboratory, reading the instruments through its cameras. Evaluation checks what the agent came to know, grounded in readings recorded during the episode, as well as the physical state the task requires; every experiment is verified to be solvable. We evaluate six frontier models in their developers' coding agents with native observation and control, with added perception and motion planning, and with manipulation skills distilled from the strongest agents. With native control, even the strongest agent succeeds less than half the time; for every agent, most failures arise in interacting with the apparatus, and manipulation skills help weaker agents more than perception and planning. The benchmark and its evaluation code are available at https://github.com/HexaFuture/EmbodiedLab.
User memory helps personalized systems build on their understanding of users over time. In e-commerce, users convey their preferences and needs through both explicit language and implicit interactions with products. We study memory built from logs of these interactions, examining how it can \emph{remember} concrete evidence and established interests, and use that evidence to \emph{infer} underlying needs and preferences that inform judgments beyond well-observed product domains. Existing memory methods primarily target language-based interactions, and evaluations offer limited real behavioral data and limited separation of these capabilities. This leaves open how to jointly support and separately evaluate remembering and inference from behavioral logs. We formulate the problem of continual behavioral user memory around these joint requirements and develop a dedicated evaluation framework and memory method to address them. \textbf{BehaviorMemBench} is a benchmark built from real e-commerce logs covering 1{,}194 users and 17{,}462 behavioral batches over more than three years, with automatic future-behavior labels and a task-independent memory interface to separately evaluate remembering and inference under a common protocol. \textbf{BehaviorMem} is a memory design that organizes explicit behavioral facts and inferred user characteristics across long-term and short-term scopes in a two-axis state, with evidence-grounded formation, continual maintenance, and query-independent readout. Experiments across two builders and two task readers show that BehaviorMem leads the evaluated agentic memories on remember and infer, and reveal different method rankings across the two capabilities and competitive statistical baselines. Validation on a public dataset further supports its usefulness for downstream tasks based on behavioral histories.
Most action-conditioned video world models remain limited to a single player, while existing multiplayer models support only small, fixed groups of agents. We introduce FlockWorld, an artificial-life environment built on Reynolds's Boids algorithm for studying population-scale multi-agent video world models. Specifically, FlockWorld renders synchronized egocentric crops, assigns distinct colors to preserve camera-agent identity across views, exposes rule-derived accelerations as actions, and lets the number of camera agents be configured independently of the size of the surrounding population. We further introduce FlockDiT, an action-conditioned latent diffusion transformer trained with flow matching on FlockWorld data to jointly generate synchronized egocentric videos for multiple agents. Our results suggest that joint generation benefits agent detection at density-matched settings and that diffusion forcing improves rollout fidelity. Together, FlockWorld and FlockDiT provide a controlled testbed for multi-agent video generation.
Junzhe Wang, Zhiheng Xi, Yajie Yang, Hao Luo, Shihan Dou, Tao Gui Fudan University, Qi Zhang Fudan University
Search agents extend Large Language Models (LLMs) beyond static parametric knowledge by enabling access to up-to-date and long-tail information unavailable during pretraining. While reinforcement learning has been widely adopted for training such agents, existing approaches face key limitations: process supervision often suffers from unstable value estimation, whereas outcome supervision struggles with credit assignment due to sparse, trajectory-level rewards. To bridge this gap, we propose Contribution-Weighted GRPO (CW-GRPO), a framework that integrates process supervision into group relative policy optimization. Instead of directly optimizing process rewards, CW-GRPO employs an LLM judge to assess the retrieval utility and reasoning correctness at each search round, producing per-round contribution weights. These weights are used to rescale outcome-based advantages along the trajectory, enabling fine-grained credit assignment without sacrificing optimization stability. Experiments on multiple knowledge-intensive benchmarks show that CW-GRPO outperforms standard GRPO by 5.0% on Qwen3-8B and 6.3% on Qwen3-1.7B, leading to more effective search behaviors. Additional analysis reveals that successful trajectories exhibit concentrated contributions in specific rounds, providing empirical insight into search agent tasks. Our code is available at https://github.com/zsxmwjz/CW-GRPO.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Sizhe Wang Washington University in St. Louis, Ziqi Xu, Claire Najjuuko Washington University, Saint Louis, Charles Alba, Chenyang Lu Washington University, Saint Louis
Clinical language models (LMs) are increasingly applied to support clinical risk prediction from free-text notes, yet their uncertainty estimates often remain poorly calibrated and clinically unreliable. In this work, we propose Clinical Uncertainty Risk Alignment (CURA), a framework that aligns clinical LM-based risk estimates and uncertainty with both individual error likelihoods and cohort-level ambiguities. CURA first fine-tunes domain-specific clinical LMs to obtain task-adapted patient embeddings, and then performs uncertainty fine-tuning of a multi-head classifier using a bi-level uncertainty objective. Specifically, an individual-level calibration term aligns predictive uncertainty with each patient’s likelihood of error, while a cohort-aware regularizer pulls risk estimates toward event rates in their local neighborhoods in the embedding space and places extra weight on ambiguous cohorts near the decision boundary. We further show that this cohort-aware term can be interpreted as a cross-entropy loss with neighborhood-informed soft labels, providing a label-smoothing view of our method. Extensive experiments on MIMIC-IV clinical risk prediction tasks across various clinical LMs show that CURA consistently improves calibration metrics without substantially compromising discrimination. Further analysis illustrates that CURA reduces overconfident false reassurance and yields more trustworthy uncertainty estimates for downstream clinical decision support.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Yunzhe Wang University of Southern California, Runhui Xu University of Southern California, Kexin Zheng University of Southern California, Tianyi Zhang, Jayavibhav Niranjan Kogundi, Soham Hans University of Southern California, Volkan Ustun USC Institute for Creative Technologies, University of Southern California
Multimodal LLMs are increasingly deployed as perceptual backbones for autonomous agents in 3D environments, from robotics to virtual worlds. These applications require agents to perceive rapid state changes, attribute actions to the correct entities, and reason about concurrent multi-agent behaviors from a first-person perspective, capabilities that existing benchmarks do not adequately evaluate. We introduce GameplayQA, a framework for evaluating agentic-centric perception and reasoning through video understanding. Specifically, we densely annotate multiplayer 3D gameplay videos at 1.22 labels/second, with time-synced, concurrent captions of states, actions, and events structured around a triadic system of Self, Other Agents, and the World, a natural decomposition for multi-agent environments. From these annotations, we refined 2.4K diagnostic QA pairs organized into three levels of cognitive complexity, accompanied by a structured distractor taxonomy that enables fine-grained analysis of where models hallucinate. Evaluation of frontier MLLMs reveals a substantial gap from human performance, with common failures in temporal and cross-video grounding, agent-role attribution, and handling the decision density of the game. We hope GameplayQA stimulates future research at the intersection of embodied AI, agentic perception, and world modeling.
尚无 AI 解读，展示原文摘要
cs.CL · cs.AI · cs.MA · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Qin Zhou East China University of Science and Technology, Guoyan Liang, Qianyi Yang, Jingyuan Chen Zhejiang University, Sai Wu Zhejiang University, Chang Yao Zhejiang University, Zhe Wang
Recent reinforcement learning (RL) approaches have advanced radiology report generation (RRG), yet two core limitations persist: (1) report-level rewards offer limited evidence-grounded guidance for clinical faithfulness; and (2) current methods lack an explicit self-improving mechanism to align with clinical preference. We introduce clinically aligned Evidence-aware Self-Correcting Reinforcement Learning (ESC-RL), comprising two key components. First, a Group-wise Evidence-aware Alignment Reward (GEAR) delivers group-wise, evidence-aware feedback. GEAR reinforces consistent grounding for true positives, recovers missed findings for false negatives, and suppresses unsupported content for false positives. Second, a Self-correcting Preference Learning (SPL) strategy automatically constructs a reliable, disease-aware preference dataset from multiple noisy observations and leverages an LLM to synthesize refined reports without human supervision. ESC-RL promotes clinically faithful, disease-aligned reward and supports continual self-improvement during training. Extensive experiments on two public chest X-ray datasets demonstrate consistent gains and state-of-the-art performance.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Jingwen Pu University of Electronic Science and Technology of China, Mingjun Shi, Xinrui Ren University of Electronic Science and Technology of China, Yizhe Wang, Xinyu Zhang, Zhaokun Wang, Kun She University of Electronic Science and Technology of China
Large Language Models (LLMs) exhibit powerful capabilities but inevitably memorize sensitive information, raising privacy, copyright, and safety concerns. Existing LLM unlearning methods typically rely on updating model parameters. While effective, they are often limited in real-world scenarios: fine-tuning large-scale models is costly, may introduce potential irreversible risks, and depends on both forget and retain datasets, which are often difficult to obtain in full. To address these challenges, an ideal solution is to achieve unlearning at inference time. To this end, we propose SEGUE, a training-free, plug-and-play inference-time unlearning strategy. SEGUE employs a probe to detect queries involving forgettable concepts and applies entropy-guided decoding to suppress target knowledge, enabling controllable non-factual generation while preserving overall model capabilities. Experiments on the MUSE, RWKU, and WMDP datasets, covering copyright, entity, and potential-risk knowledge, show that SEGUE effectively balances sensitive knowledge suppression and generation quality, outperforming existing most inference-time unlearning methods.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Sizhe Wang, Zhengren Wang Peking University, Dongsheng Ma, Yongan Yu Mila - Quebec Artificial Intelligence Institute and McGill University, McGill University, Rui Ling, Zhiyu li, Feiyu Xiong MemTensor (Shanghai) Technology Co., Ltd., Wentao Zhang
Modern software development demands code that is maintainable, testable, and scalable by organizing the implementation into modular components with iterative reuse of existing codes. We formalize this iterative, multi-turn paradigm as codeflow and introduce CodeFlowBench, the first benchmark designed to comprehensively evaluate LLMs’ ability to perform codeflow - implementing new functionality by reusing existing functions over multiple turns. CodeFlowBench comprises two complementary components: CodeFlowBench-Comp, a core collection of 5,000+ competitive programming problems from Codeforces updated via an automated pipeline and CodeFlowBench-Repo, which is sourced from GitHub repositories to better reflect real-world scenarios. Furthermore, a novel evaluation framework featured dual assessment protocol and structural metrics derived from dependency trees is introduced. Extensive experiments reveal significant performance degradation in multi-turn codeflow scenarios. Furthermore, our in-depth analysis illustrates that model performance inversely correlates with dependency complexity. These findings not only highlight the critical challenges for supporting real-world workflows, but also establish CodeFlowBench as an essential tool for advancing code generation research.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Language agents, i.e., LLM agents, progress rapidly and are increasingly deployed in production environments. This trend underscores the urgent need for rigorous and realistic evaluations. However, most existing benchmarks evaluate agents in simplified, idealized settings. They typically rely on pre-packaged tool interfaces, overlook critical steps, and assume inputs are clean and fully specified. Consequently, they understate the difficulty of real deployments, where uncertainty and noise are ubiquitous and agents must proactively explore the environment to uncover new tools. To bridge this gap, we present AgentGym2, a new evaluation framework with task instances grounded in real-world end-to-end working demands. Beyond reasoning and planning, it measures agents’ ability to execute end-to-end procedures, discover tools via exploration, compose tools for unseen tasks, and remain robust to noisy and underspecified information. Experiments on 15 proprietary and open-source models show that even SOTA systems like Gemini and GPT-5 struggle on AgentGym2, revealing a substantial gap between the capability of current agents and the demands of real-world applications.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Ming Zhang Fudan University, Yujiong Shen, Jingyi Deng, Yuhui Wang, Huayu Sha, Kexin Tan The University of Tokyo, The University of Tokyo and Fudan University, Qiyuan Peng, Yue Zhang ATOM INFINITE PTE. LTD. 等 22 位作者
Existing evaluation of Large Language Models (LLMs) on static benchmarks is vulnerable to data contamination and leaderboard overfitting, critical issues that obscure true model capabilities. To address this, we introduce LLMEval-Fair, a framework for dynamic evaluation of LLMs. LLMEval-Fair is built on a proprietary bank of 220k graduate-level questions, from which it dynamically samples unseen test sets for each evaluation run. Its automated pipeline ensures integrity via contamination-resistant data curation, a novel anti-cheating architecture, and a calibrated LLM-as-a-judge process achieving 90% agreement with human experts, complemented by a relative ranking system for fair comparison. An 30-month longitudinal study of nearly 60 leading models reveals a performance ceiling on knowledge memorization and exposes data contamination vulnerabilities undetectable by static benchmarks. The framework demonstrates exceptional robustness in ranking stability and consistency, providing strong empirical validation for the dynamic evaluation paradigm. LLMEval-Fair offers a robust and credible methodology for assessing the true capabilities of LLMs beyond leaderboard scores, promoting the development of more trustworthy evaluation standards.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表