Daily Papers

Large language models (LLMs) have demonstrated excellent capabilities in generating structured diagrams from natural language descriptions. In particular, they have shown great promise in generating sequence diagrams for software engineering, typically represented in a text-based syntax such as Mermaid. However, systematic evaluations in this space remain underdeveloped as there is a lack of existing benchmarks to assess the LLM's correctness in this task. To address this shortcoming, we introduce MermaidSeqBench, a human-verified and LLM-synthetically-extended benchmark for assessing an LLM's capabilities in generating Mermaid sequence diagrams from textual prompts. The benchmark consists of a core set of 132 samples, starting from a small set of manually crafted and verified flows. These were expanded via a hybrid methodology combining human annotation, in-context LLM prompting, and rule-based variation generation. Our benchmark uses an LLM-as-a-judge model to assess Mermaid sequence diagram generation across fine-grained metrics, including syntax correctness, activation handling, error handling, and practical usability. We perform initial evaluations on numerous state-of-the-art LLMs and utilize multiple LLM judge models to demonstrate the effectiveness and flexibility of our benchmark. Our results reveal significant capability gaps across models and evaluation modes. Our proposed benchmark provides a foundation for advancing research in structured diagram generation and for developing more rigorous, fine-grained evaluation methodologies.

Benchmarks are crucial for evaluating progress in robotics and embodied AI. However, a significant gap exists between benchmarks designed for high-level language instruction following, which often assume perfect low-level execution, and those for low-level robot control, which rely on simple, one-step commands. This disconnect prevents a comprehensive evaluation of integrated systems where both task planning and physical execution are critical. To address this, we propose Kitchen-R, a novel benchmark that unifies the evaluation of task planning and low-level control within a simulated kitchen environment. Built as a digital twin using the Isaac Sim simulator and featuring more than 500 complex language instructions, Kitchen-R supports a mobile manipulator robot. We provide baseline methods for our benchmark, including a task-planning strategy based on a vision-language model and a low-level control policy based on diffusion policy. We also provide a trajectory collection system. Our benchmark offers a flexible framework for three evaluation modes: independent assessment of the planning module, independent assessment of the control policy, and, crucially, an integrated evaluation of the whole system. Kitchen-R bridges a key gap in embodied AI research, enabling more holistic and realistic benchmarking of language-guided robotic agents.

Deep Research Agents are increasingly used for automated survey generation. However, whether they can write surveys like human experts remains unclear. Existing benchmarks focus on fluency or citation accuracy, but none evaluates the core capabilities: retrieving essential papers and organizing them into coherent knowledge structures. We introduce TaxoBench, a diagnostic benchmark derived from 72 highly-cited computer science surveys. We manually extract expert-authored taxonomy trees containing 3,815 precisely categorized citations as ground truth. Our benchmark supports two evaluation modes: Deep Research mode tests end-to-end retrieval and organization given only a topic, while Bottom-Up mode isolates structuring capability by providing the exact papers human experts used. We evaluate 7 leading Deep Research agents and 12 frontier LLMs. Results reveal a dual bottleneck: the best agent recalls only 20.9% of expert-selected papers, and even with perfect input, the best model achieves only 0.31 ARI in organization. Current deep research agents remain far from expert-level survey writing. Our benchmark is publicly available at https://github.com/KongLongGeFDU/TaxoBench.

While LLM-Based agents, which use external tools to solve complex problems, have made significant progress, benchmarking their ability is challenging, thereby hindering a clear understanding of their limitations. In this paper, we propose an interactive evaluation framework, named CIBench, to comprehensively assess LLMs' ability to utilize code interpreters for data science tasks. Our evaluation framework includes an evaluation dataset and two evaluation modes. The evaluation dataset is constructed using an LLM-human cooperative approach and simulates an authentic workflow by leveraging consecutive and interactive IPython sessions. The two evaluation modes assess LLMs' ability with and without human assistance. We conduct extensive experiments to analyze the ability of 24 LLMs on CIBench and provide valuable insights for future LLMs in code interpreter utilization.

Multimodal Large Language Models (MLLMs) have shown significant potential in medical image analysis. However, their capabilities in interpreting fundus images, a critical skill for ophthalmology, remain under-evaluated. Existing benchmarks lack fine-grained task divisions and fail to provide modular analysis of its two key modules, i.e., large language model (LLM) and vision encoder (VE). This paper introduces FunBench, a novel visual question answering (VQA) benchmark designed to comprehensively evaluate MLLMs' fundus reading skills. FunBench features a hierarchical task organization across four levels (modality perception, anatomy perception, lesion analysis, and disease diagnosis). It also offers three targeted evaluation modes: linear-probe based VE evaluation, knowledge-prompted LLM evaluation, and holistic evaluation. Experiments on nine open-source MLLMs plus GPT-4o reveal significant deficiencies in fundus reading skills, particularly in basic tasks such as laterality recognition. The results highlight the limitations of current MLLMs and emphasize the need for domain-specific training and improved LLMs and VEs.

The rapid evolution of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has produced substantial gains in reasoning, perception, and generative capability across language and vision. However, whether these advances yield commensurate improvements in safety remains unclear, in part due to fragmented evaluation practices limited to single modalities or threat models. In this report, we present an integrated safety evaluation of 7 frontier models: GPT-5.2, Gemini 3 Pro, Qwen3-VL, Doubao 1.8, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5. We evaluate each model across language, vision-language, and image generation settings using a unified protocol that integrates benchmark evaluation, adversarial evaluation, multilingual evaluation, and compliance evaluation. Aggregating our evaluations into safety leaderboards and model safety profiles across multiple evaluation modes reveals a sharply heterogeneous safety landscape. While GPT-5.2 demonstrates consistently strong and balanced safety performance across evaluations, other models exhibit pronounced trade-offs among benchmark safety, adversarial alignment, multilingual generalization, and regulatory compliance. Both language and vision-language modalities show significant vulnerability under adversarial evaluation, with all models degrading substantially despite strong results on standard benchmarks. Text-to-image models achieve relatively stronger alignment in regulated visual risk categories, yet remain brittle under adversarial or semantically ambiguous prompts. Overall, these results show that safety in frontier models is inherently multidimensional--shaped by modality, language, and evaluation scheme, underscoring the need for standardized safety evaluations to accurately assess real-world risk and guide responsible model development and deployment.

Large Language Models are being increasingly deployed as the decision-making core of autonomous agents capable of effecting change in external environments. Yet, in conversational benchmarks, which simulate real-world customer-centric issue resolution scenarios, these agents frequently fail due to the cascading effects of incorrect decision-making. These challenges are particularly pronounced for open-source LLMs with smaller parameter sizes, limited context windows, and constrained inference budgets, which contribute to increased error accumulation in agentic settings. To tackle these challenges, we present the Failure-Aware Meta-Agentic (FAMA) framework. FAMA operates in two stages: first, it analyzes failure trajectories from baseline agents to identify the most prevalent errors; second, it employs an orchestration mechanism that activates a minimal subset of specialized agents tailored to address these failures by injecting a targeted context for the tool-use agent before the decision-making step. Experiments across open-source LLMs demonstrate performance gains up to 27% across evaluation modes over standard baselines. These results highlight that targeted curation of context through specialized agents to address common failures is a valuable design principle for building reliable, multi-turn tool-use LLM agents that simulate real-world conversational scenarios.

Recently, multimodal large language models (MLLMs) have achieved significant advancements across various domains, and corresponding evaluation benchmarks have been continuously refined and improved. In this process, benchmarks in the scientific domain have played an important role in assessing the reasoning capabilities of MLLMs. However, existing benchmarks still face three key challenges: 1) Insufficient evaluation of models' reasoning abilities in multilingual scenarios; 2) Inadequate assessment of MLLMs' comprehensive modality coverage; 3) Lack of fine-grained annotation of scientific knowledge points. To address these gaps, we propose MME-SCI, a comprehensive and challenging benchmark. We carefully collected 1,019 high-quality question-answer pairs, which involve 3 distinct evaluation modes. These pairs cover four subjects, namely mathematics, physics, chemistry, and biology, and support five languages: Chinese, English, French, Spanish, and Japanese. We conducted extensive experiments on 16 open-source models and 4 closed-source models, and the results demonstrate that MME-SCI is widely challenging for existing MLLMs. For instance, under the Image-only evaluation mode, o4-mini achieved accuracy of only 52.11%, 24.73%, 36.57%, and 29.80% in mathematics, physics, chemistry, and biology, respectively, indicating a significantly higher difficulty level compared to existing benchmarks. More importantly, using MME-SCI's multilingual and fine-grained knowledge attributes, we analyzed existing models' performance in depth and identified their weaknesses in specific domains. The Data and Evaluation Code are available at https://github.com/JCruan519/MME-SCI.

Existing LLM-as-a-Judge systems suffer from three fundamental limitations: limited adaptivity to task- and domain-specific evaluation criteria, systematic biases driven by non-semantic cues such as position, length, format, and model provenance, and evaluation inconsistency that leads to contradictory judgments across different evaluation modes (e.g., pointwise versus pairwise). To address these issues, we propose FairJudge, an adaptive, debiased, and consistent LLM-as-a-Judge. Unlike prior approaches that treat the judge as a static evaluator, FairJudge models judging behavior itself as a learnable and regularized policy. From a data-centric perspective, we construct a high-information-density judging dataset that explicitly injects supervision signals aligned with evaluation behavior. Building on this dataset, we adopt a curriculum-style SFT-DPO-GRPO training paradigm that progressively aligns rubric adherence, bias mitigation, and cross-mode consistency, while avoiding catastrophic forgetting. Experimental results on multiple internal and public benchmarks show that FairJudge consistently improves agreement and F1, reduces non-semantic biases, and outperforms substantially larger instruction-tuned LLMs. All resources will be publicly released after acceptance to facilitate future research.

Federated and decentralized machine learning leverage end-user devices for privacy-preserving training of models at lower operating costs than within a data center. In a round of Federated Learning (FL), a random sample of participants trains locally, then a central server aggregates the local models to produce a single model for the next round. In a round of Decentralized Learning (DL), all participants train locally and then aggregate with their immediate neighbors, resulting in many local models with residual variance between them. On the one hand, FL's sampling and lower model variance provides lower communication costs and faster convergence. On the other hand, DL removes the need for a central server and distributes the communication costs more evenly amongst nodes, albeit at a larger total communication cost and slower convergence. In this paper, we present MoDeST: Mostly-Consistent Decentralized Sampling Training. MoDeST implements decentralized sampling in which a random subset of nodes is responsible for training and aggregation every round: this provides the benefits of both FL and DL without their traditional drawbacks. Our evaluation of MoDeST on four common learning tasks: (i) confirms convergence as fast as FL, (ii) shows a 3x-14x reduction in communication costs compared to DL, and (iii) demonstrates that MoDeST quickly adapts to nodes joining, leaving, or failing, even when 80% of all nodes become unresponsive.

As LLMs increasingly mediate the political information citizens rely on, there is still no standardized way to assess whether they do so responsibly. We introduce Polistemics, a theory-grounded benchmark for evaluating LLMs as mediators of political information in elections. Prior work has treated this task as reproduction rather than mediation, leaving its epistemic dimensions and interaction with imperfect information unaddressed. We ground the evaluation in Epistemic Modesty, a normative standard derived from citizens' epistemic agency, and test it across controlled settings that vary informational properties such as clarity, noise, and consistency. Applying the benchmark to three state-of-the-art LLMs on the 2025 German and Dutch elections, we find that high aggregate scores mask systematic failures. Models mediate reliably under clear evidence but break down under absent, vague, or contradictory information, while flattening the intensity of political language. These failures are likely driven by party priors, influenced by party labels and output language. Reliable mediation appears achievable, but no model delivers it consistently.

Large language model-based web agents have demonstrated strong performance on realistic web interaction tasks. However, existing evaluations are predominantly conducted under relatively stable and well-behaved interaction conditions, which may overestimate agent robustness. High task success in such idealized settings does not necessarily reflect performance under realistic web interaction. To address this limitation, we introduce a diagnostic stress-testing benchmark for web agents. We first construct realistic and controllable web environments that provide clean and stable interaction workflows as reference baselines. We then introduce structured and controlled perturbations that emulate interaction variability, including shifting layouts, altered interaction semantics, and execution disruptions. By comparing agent behavior between clean and perturbed settings, our framework enables systematic diagnosis of robustness under what-if interaction scenarios. Through extensive evaluation of state-of-the-art multimodal web agents, we show that stress-based evaluation exposes failure modes and substantial robustness gaps that remain hidden under clean benchmark conditions.

Natural language is an appealing medium for explaining how large language models process and store information, but evaluating the faithfulness of such explanations is challenging. To help address this, we develop two modes of evaluation for natural language explanations that claim individual neurons represent a concept in a text input. In the observational mode, we evaluate claims that a neuron a activates on all and only input strings that refer to a concept picked out by the proposed explanation E. In the intervention mode, we construe E as a claim that the neuron a is a causal mediator of the concept denoted by E. We apply our framework to the GPT-4-generated explanations of GPT-2 XL neurons of Bills et al. (2023) and show that even the most confident explanations have high error rates and little to no causal efficacy. We close the paper by critically assessing whether natural language is a good choice for explanations and whether neurons are the best level of analysis.

Existing evaluation frameworks for large language models -- including HELM, MT-Bench, AgentBench, and BIG-bench -- are designed for controlled, single-session, lab-scale settings. They do not address the evaluation challenges that emerge when agentic AI systems operate continuously in production: compounding decision errors, tool failure cascades, non-deterministic output drift, and the absence of ground truth for long-horizon tasks. This paper makes three contributions. First, we present a taxonomy of seven failure modes unique to production agentic systems, each grounded in observations from systems operating at billion-event scale. Second, we demonstrate empirically where standard metrics -- ROUGE, BERTScore, accuracy/AUC, and the agentic benchmarks above -- fail to detect each failure mode. Third, we propose PAEF (Production Agentic Evaluation Framework), a five-dimension evaluation framework with an open-source reference implementation, designed for continuous evaluation on production traffic rather than episodic benchmark runs. Our analysis shows that standard metrics fail to detect four of the seven failure modes entirely and detect three others only after a lag of multiple evaluation cycles.

Neural PDE solvers are increasingly used as learned surrogates for families of partial differential equations, where the key machine learning challenge is not only interpolation on a fixed benchmark distribution but generalization under structured shifts in coefficients, boundary conditions, discretization, and rollout horizon. Yet evaluation is still often dominated by in-distribution test error, making robustness difficult to assess. We introduce a standardized stress-testing framework for neural PDE solvers under deployment-relevant shift. We instantiate it on three representative architectures -- Fourier Neural Operators (FNOs), a DeepONet-style model, and convolutional neural operators (CNOs) -- across five qualitatively different PDE families: dispersive, elliptic, multi-scale fluid, financial, and chaotic systems. Across 750 trained models, we measure robustness using baseline-normalized degradation factors together with spectral and rollout diagnostics. The resulting comparisons reveal that strong in-distribution accuracy does not reliably predict robustness, and that failure patterns depend jointly on architecture and PDE family. Our results provide a clearer basis for evaluating robustness claims in neural PDE solvers and suggest that function-space generalization under structured shift should be treated as a first-class evaluation target.

The financial domain poses unique challenges for knowledge graph (KG) construction at scale due to the complexity and regulatory nature of financial documents. Despite the critical importance of structured financial knowledge, the field lacks large-scale, open-source datasets capturing rich semantic relationships from corporate disclosures. We introduce an open-source, large-scale financial knowledge graph dataset built from the latest annual SEC 10-K filings of all S and P 100 companies - a comprehensive resource designed to catalyze research in financial AI. We propose a robust and generalizable knowledge graph (KG) construction framework that integrates intelligent document parsing, table-aware chunking, and schema-guided iterative extraction with a reflection-driven feedback loop. Our system incorporates a comprehensive evaluation pipeline, combining rule-based checks, statistical validation, and LLM-as-a-Judge assessments to holistically measure extraction quality. We support three extraction modes - single-pass, multi-pass, and reflection-agent-based - allowing flexible trade-offs between efficiency, accuracy, and reliability based on user requirements. Empirical evaluations demonstrate that the reflection-agent-based mode consistently achieves the best balance, attaining a 64.8 percent compliance score against all rule-based policies (CheckRules) and outperforming baseline methods (single-pass and multi-pass) across key metrics such as precision, comprehensiveness, and relevance in LLM-guided evaluations.

We introduce iTRIALSPACE, a programmable evaluation framework for controlled assessment of lung CT models. Standard benchmarks are static retrospective collections that entangle lesion size, lobe prevalence, anatomy, and acquisition context, making it difficult to determine what structurally drives model accuracy. iTRIALSPACE addresses this limitation by composing real clinical CTs and lesion profiles into controlled virtual lesion trials through a four-stage pipeline: multidataset nodule profiling, explicit trial specification, anatomy-aware mask insertion, and ControlNet-conditioned CT synthesis. The framework is built on a unified 54-attribute nodule-profile dataset spanning 13,140 annotated nodules from seven public CT sources and instantiated as 13 trial modes. We evaluate iTRIALSPACE in a 55,469-sample Virtual Lesion Study spanning three medical VLMs, four spatialguidance conditions, and three clinical tasks. Across all 13 modes, the synthetic substrate remains within the real-to-real FID baseline, and synthetic performance rankings transfer strongly to real clinical data (ρ = 0.93, p < 10^{-15}). Controlled trial modes expose findings unavailable to fixed-distribution benchmarks, including shortcut-driven size prediction collapse under lobe-equalized sampling and hostto-donor variance ratios of 8.9x and 3.3x in twin-cross analysis. These results position iTRIALSPACE as an auditable evaluation infrastructure for controlled, falsifiable testing beyond static retrospective benchmarks.

We present the analytic evaluation of the gravitational energy and of the angular momentum flux with tidal effects for inspiraling compact binaries, at next-to-next-to-leading post-Newtoian (2PN) order, within the effective field theory diagrammatic approach. We first compute the stress-energy tensor for a binary system, that requires the evaluation of two-point Feynman integrals, up to two loops. Then, we extract the multipole moments of the system, which we present for generic orbits in center-of-mass coordinates, and which are needed for the evaluation of the total gravitational energy and the angular momentum flux, for generic orbits. Finally, we provide the expression of gauge invariant quantities such as the fluxes, and the mode amplitudes and phase of the emitted gravitational wave, for circular orbits. Our findings are useful to update earlier theoretical studies as well as related phenomenological analyses, and waveform models

Recent advances in coding agents suggest rapid progress toward autonomous software development, yet existing benchmarks fail to rigorously evaluate the long-horizon capabilities required to build complete software systems. Most prior evaluations focus on localized code generation, scaffolded completion, or short-term repair tasks, leaving open the question of whether agents can sustain coherent reasoning, planning, and execution over the extended horizons demanded by real-world repository construction. To address this gap, we present NL2Repo Bench, a benchmark explicitly designed to evaluate the long-horizon repository generation ability of coding agents. Given only a single natural-language requirements document and an empty workspace, agents must autonomously design the architecture, manage dependencies, implement multi-module logic, and produce a fully installable Python library. Our experiments across state-of-the-art open- and closed-source models reveal that long-horizon repository generation remains largely unsolved: even the strongest agents achieve below 40% average test pass rates and rarely complete an entire repository correctly. Detailed analysis uncovers fundamental long-horizon failure modes, including premature termination, loss of global coherence, fragile cross-file dependencies, and inadequate planning over hundreds of interaction steps. NL2Repo Bench establishes a rigorous, verifiable testbed for measuring sustained agentic competence and highlights long-horizon reasoning as a central bottleneck for the next generation of autonomous coding agents.

Interpretation is essential to deciphering the language of art: audiences communicate with artists by recovering meaning from visual artifacts. However, current Generative Art (GenArt) evaluators remain fixated on surface-level image quality or literal prompt adherence, failing to assess the deeper symbolic or abstract meaning intended by the creator. We address this gap by formalizing a Peircean computational semiotic theory that models Human-GenArt Interaction (HGI) as cascaded semiosis. This framework reveals that artistic meaning is conveyed through three modes - iconic, symbolic, and indexical - yet existing evaluators operate heavily within the iconic mode, remaining structurally blind to the latter two. To overcome this structural blindness, we propose SemJudge. This evaluator explicitly assesses symbolic and indexical meaning in HGI via a Hierarchical Semiosis Graph (HSG) that reconstructs the meaning-making process from prompt to generated artifact. Extensive quantitative experiments show that SemJudge aligns more closely with human judgments than prior evaluators on an interpretation-intensive fine-art benchmark. User studies further demonstrate that SemJudge produces deeper, more insightful artistic interpretations, thereby paving the way for GenArt to move beyond the generation of "pretty" images toward a medium capable of expressing complex human experience. Project page: https://github.com/songrise/SemJudge.

Optimizing communication topology in LLM-based multi-agent system is critical for enabling collective intelligence. Existing methods mainly rely on spatio-temporal interaction paradigms, where the sequential execution of multi-round dialogues incurs high latency and computation. Motivated by the recent insights that evaluation and debate mechanisms can improve problem-solving in multi-agent systems, we propose TopoDIM, a framework for one-shot Topology generation with Diverse Interaction Modes. Designed for decentralized execution to enhance adaptability and privacy, TopoDIM enables agents to autonomously construct heterogeneous communication without iterative coordination, achieving token efficiency and improved task performance. Experiments demonstrate that TopoDIM reduces total token consumption by 46.41% while improving average performance by 1.50% over state-of-the-art methods. Moreover, the framework exhibits strong adaptability in organizing communication among heterogeneous agents. Code is available at: https://anonymous.4open.science/r/TopoDIM-8D35/

Objective: BCI (Brain-Computer Interface) technology operates in three modes: online, offline, and pseudo-online. In the online mode, real-time EEG data is constantly analyzed. In offline mode, the signal is acquired and processed afterwards. The pseudo-online mode processes collected data as if they were received in real-time. The main difference is that the offline mode often analyzes the whole data, while the online and pseudo-online modes only analyze data in short time windows. Offline analysis is usually done with asynchronous BCIs, which restricts analysis to predefined time windows. Asynchronous BCI, compatible with online and pseudo-online modes, allows flexible mental activity duration. Offline processing tends to be more accurate, while online analysis is better for therapeutic applications. Pseudo-online implementation approximates online processing without real-time constraints. Many BCI studies being offline introduce biases compared to real-life scenarios, impacting classification algorithm performance. Approach: The objective of this research paper is therefore to extend the current MOABB framework, operating in offline mode, so as to allow a comparison of different algorithms in a pseudo-online setting with the use of a technology based on overlapping sliding windows. To do this will require the introduction of a idle state event in the dataset that takes into account all different possibilities that are not task thinking. To validate the performance of the algorithms we will use the normalized Matthews Correlation Coefficient (nMCC) and the Information Transfer Rate (ITR). Main results: We analyzed the state-of-the-art algorithms of the last 15 years over several Motor Imagery (MI) datasets composed by several subjects, showing the differences between the two approaches from a statistical point of view. Significance: The ability to analyze the performance of different algorithms in offline and pseudo-online modes will allow the BCI community to obtain more accurate and comprehensive reports regarding the performance of classification algorithms.

Interests in the automatic generation of cooking recipes have been growing steadily over the past few years thanks to a large amount of online cooking recipes. We present RecipeGPT, a novel online recipe generation and evaluation system. The system provides two modes of text generations: (1) instruction generation from given recipe title and ingredients; and (2) ingredient generation from recipe title and cooking instructions. Its back-end text generation module comprises a generative pre-trained language model GPT-2 fine-tuned on a large cooking recipe dataset. Moreover, the recipe evaluation module allows the users to conveniently inspect the quality of the generated recipe contents and store the results for future reference. RecipeGPT can be accessed online at https://recipegpt.org/.

As intelligent agents become more generally-capable, i.e. able to master a wide variety of tasks, the complexity and cost of properly evaluating them rises significantly. Tasks that assess specific capabilities of the agents can be correlated and stochastic, requiring many samples for accurate comparisons, leading to added costs. In this paper, we propose a formal definition and a conceptual framework for active evaluation of agents across multiple tasks, which assesses the performance of ranking algorithms as a function of number of evaluation data samples. Rather than curating, filtering, or compressing existing data sets as a preprocessing step, we propose an online framing: on every iteration, the ranking algorithm chooses the task and agents to sample scores from. Then, evaluation algorithms report a ranking of agents on each iteration and their performance is assessed with respect to the ground truth ranking over time. Several baselines are compared under different experimental contexts, with synthetic generated data and simulated online access to real evaluation data from Atari game-playing agents. We find that the classical Elo rating system -- while it suffers from well-known failure modes, in theory -- is a consistently reliable choice for efficient reduction of ranking error in practice. A recently-proposed method, Soft Condorcet Optimization, shows comparable performance to Elo on synthetic data and significantly outperforms Elo on real Atari agent evaluation. When task variation from the ground truth is high, selecting tasks based on proportional representation leads to higher rate of ranking error reduction.

Large language models (LLMs) are being rapidly integrated into decision-support tools, automation workflows, and AI-enabled software systems. However, their behavior in production environments remains poorly understood, and their failure patterns differ fundamentally from those of traditional machine learning models. This paper presents a system-level taxonomy of fifteen hidden failure modes that arise in real-world LLM applications, including multi-step reasoning drift, latent inconsistency, context-boundary degradation, incorrect tool invocation, version drift, and cost-driven performance collapse. Using this taxonomy, we analyze the growing gap in evaluation and monitoring practices: existing benchmarks measure knowledge or reasoning but provide little insight into stability, reproducibility, drift, or workflow integration. We further examine the production challenges associated with deploying LLMs - including observability limitations, cost constraints, and update-induced regressions - and outline high-level design principles for building reliable, maintainable, and cost-aware LLM systems. Finally, we outline high-level design principles for building reliable, maintainable, and cost-aware LLM-based systems. By framing LLM reliability as a system-engineering problem rather than a purely model-centric one, this work provides an analytical foundation for future research on evaluation methodology, AI system robustness, and dependable LLM deployment.

Current evaluations of spatial intelligence can be systematically invalid under modern vision-language model (VLM) settings. First, many benchmarks derive question-answer (QA) pairs from point-cloud-based 3D annotations originally curated for traditional 3D perception. When such annotations are treated as ground truth for video-based evaluation, reconstruction and annotation artifacts can miss objects that are clearly visible in the video, mislabel object identities, or corrupt geometry-dependent answers (e.g., size), yielding incorrect or ambiguous QA pairs. Second, evaluations often assume full-scene access, while many VLMs operate on sparsely sampled frames (e.g., 16-64), making many questions effectively unanswerable under the actual model inputs. We improve evaluation validity by introducing ReVSI, a benchmark and protocol that ensures each QA pair is answerable and correct under the model's actual inputs. To this end, we re-annotate objects and geometry across 381 scenes from 5 datasets to improve data quality, and regenerate all QA pairs with rigorous bias mitigation and human verification using professional 3D annotation tools. We further enhance evaluation controllability by providing variants across multiple frame budgets (16/32/64/all) and fine-grained object visibility metadata, enabling controlled diagnostic analyses. Evaluations of general and domain-specific VLMs on ReVSI reveal systematic failure modes that are obscured by prior benchmarks, yielding a more reliable and diagnostic assessment of spatial intelligence.

Physical interactions follow a long-tailed distribution: a set of common and regular interactions dominates human experience and visual data, while a broad spectrum of rare and irregular interactions remains underrepresented. Although recent visual world models, including image and video generation models, achieve impressive realism on existing benchmarks, they primarily focus on simulating common physical interactions. This raises a central question: Do current visual world models internalize and generalize physical principles? In this work, we introduce Tailor-Bench, a benchmark that challenges world models to simulate irregular physical interactions. To enable systematic evaluation, we design three scenario modes that progressively challenge model reasoning: Regular scenarios reflect common tool-task pairs, Unconventional scenarios replace conventional tools with attribute-compatible substitutes to test affordance generalization, and Impossible scenarios introduce attribute-violating tools to probe constraint awareness. Additionally, we design two complementary settings under a unified evaluation protocol: predictive generation requires inferring outcomes without guidance, while descriptive generation specifies the target outcome for faithful realization. Our experimental results reveal a clear long-tail gap in physical world modeling: performance degrades from Regular to Unconventional and Impossible scenarios, indicating limited generalization beyond common interactions. Failure analysis further shows that models rely on superficial visual patterns: image models fail to realize correct state changes, while video models further suffer from temporal inconsistencies.

Prompt design is a primary control interface for large language models (LLMs), yet standard evaluations largely reduce performance to answer correctness, obscuring why a prompt succeeds or fails and providing little actionable guidance. We propose PEEM (Prompt Engineering Evaluation Metrics), a unified framework for joint and interpretable evaluation of both prompts and responses. PEEM defines a structured rubric with 9 axes: 3 prompt criteria (clarity/structure, linguistic quality, fairness) and 6 response criteria (accuracy, coherence, relevance, objectivity, clarity, conciseness), and uses an LLM-based evaluator to output (i) scalar scores on a 1-5 Likert scale and (ii) criterion-specific natural-language rationales grounded in the rubric. Across 7 benchmarks and 5 task models, PEEM's accuracy axis strongly aligns with conventional accuracy while preserving model rankings (aggregate Spearman rho about 0.97, Pearson r about 0.94, p < 0.001). A multi-evaluator study with four models shows consistent relative judgments (pairwise rho = 0.68-0.85), supporting evaluator-agnostic deployment. Beyond alignment, PEEM captures complementary linguistic failure modes and remains informative under prompt perturbations: prompt-quality trends track downstream accuracy under iterative rewrites, semantic adversarial manipulations induce clear score degradation, and meaning-preserving paraphrases yield high stability (robustness rate about 76.7-80.6%). Finally, using only PEEM scores and rationales as feedback, a zero-shot prompt rewriting loop improves downstream accuracy by up to 11.7 points, outperforming supervised and RL-based prompt-optimization baselines. Overall, PEEM provides a reproducible, criterion-driven protocol that links prompt formulation to response behavior and enables systematic diagnosis and optimization of LLM interactions.

Reinforcement learning (RL), large language models (LLMs), and vision-language models (VLMs) have been widely studied in isolation. However, existing infrastructure lacks the ability to deploy agents from different decision-making paradigms within the same environment, making it difficult to study them in hybrid multi-agent settings or to compare their behaviour fairly under identical conditions. We present MOSAIC, an open-source platform that bridges this gap by incorporating a diverse set of existing reinforcement learning environments and enabling heterogeneous agents (RL policies, LLMs, VLMs, and human players) to operate within them in ad-hoc team settings with reproducible results. MOSAIC introduces three contributions. (i) An IPC-based worker protocol that wraps both native and third-party frameworks as isolated subprocess workers, each executing its native training and inference logic unmodified, communicating through a versioned inter-process protocol. (ii) An operator abstraction that forms an agent-level interface by mapping workers to agents: each operator, regardless of whether it is backed by an RL policy, an LLM, or a human, conforms to a minimal unified interface. (iii) A deterministic cross-paradigm evaluation framework offering two complementary modes: a manual mode that advances up to N concurrent operators in lock-step under shared seeds for fine-grained visual inspection of behavioural differences, and a script mode that drives automated, long-running evaluation through declarative Python scripts, for reproducible experiments. We release MOSAIC as an open, visual-first platform to facilitate reproducible cross-paradigm research across the RL, LLM, and human-in-the-loop communities.

The increasing context length of modern language models has created a need for evaluating their ability to retrieve and process information across extensive documents. While existing benchmarks test long-context capabilities, they often lack a structured way to systematically vary question complexity. We introduce KG-QAGen (Knowledge-Graph-based Question-Answer Generation), a framework that (1) extracts QA pairs at multiple complexity levels (2) by leveraging structured representations of financial agreements (3) along three key dimensions -- multi-hop retrieval, set operations, and answer plurality -- enabling fine-grained assessment of model performance across controlled difficulty levels. Using this framework, we construct a dataset of 20,139 QA pairs (the largest number among the long-context benchmarks) and open-source a part of it. We evaluate 13 proprietary and open-source LLMs and observe that even the best-performing models are struggling with set-based comparisons and multi-hop logical inference. Our analysis reveals systematic failure modes tied to semantic misinterpretation and inability to handle implicit relations.

Evaluating agentic AI on open-ended professional tasks faces a fundamental dilemma between rigor and flexibility. Static rubrics provide rigorous, reproducible assessment but fail to accommodate diverse valid response strategies, while LLM-as-a-judge approaches adapt to individual responses yet suffer from instability and bias. Human experts address this dilemma by combining domain-grounded principles with dynamic, claim-level assessment. Inspired by this process, we propose JADE, a two-layer evaluation framework. Layer 1 encodes expert knowledge as a predefined set of evaluation skills, providing stable evaluation criteria. Layer 2 performs report-specific, claim-level evaluation to flexibly assess diverse reasoning strategies, with evidence-dependency gating to invalidate conclusions built on refuted claims. Experiments on BizBench show that JADE improves evaluation stability and reveals critical agent failure modes missed by holistic LLM-based evaluators. We further demonstrate strong alignment with expert-authored rubrics and effective transfer to a medical-domain benchmark, validating JADE across professional domains. Our code is publicly available at https://github.com/smiling-world/JADE.

This paper introduces BioAgent Bench, a benchmark dataset and an evaluation suite designed for measuring the performance and robustness of AI agents in common bioinformatics tasks. The benchmark contains curated end-to-end tasks (e.g., RNA-seq, variant calling, metagenomics) with prompts that specify concrete output artifacts to support automated assessment, including stress testing under controlled perturbations. We evaluate frontier closed-source and open-weight models across multiple agent harnesses, and use an LLM-based grader to score pipeline progress and outcome validity. We find that frontier agents can complete multi-step bioinformatics pipelines without elaborate custom scaffolding, often producing the requested final artifacts reliably. However, robustness tests reveal failure modes under controlled perturbations (corrupted inputs, decoy files, and prompt bloat), indicating that correct high-level pipeline construction does not guarantee reliable step-level reasoning. Finally, because bioinformatics workflows may involve sensitive patient data, proprietary references, or unpublished IP, closed-source models can be unsuitable under strict privacy constraints; in such settings, open-weight models may be preferable despite lower completion rates. We release the dataset and evaluation suite publicly.

Audio-visual understanding is a rapidly evolving field that seeks to integrate and interpret information from both auditory and visual modalities. Despite recent advances in multi-modal learning, existing benchmarks often suffer from strong visual bias -- where answers can be inferred from visual data alone -- and provide only aggregate scores that conflate multiple sources of error. This makes it difficult to determine whether models struggle with visual understanding, audio interpretation, or audio-visual alignment. In this work, we introduce DAVE (Diagnostic Audio Visual Evaluation), a novel benchmark dataset designed to systematically evaluate audio-visual models across controlled challenges. DAVE alleviates existing limitations by (i) ensuring both modalities are necessary to answer correctly and (ii) decoupling evaluation into atomic subcategories. Our detailed analysis of state-of-the-art models reveals specific failure modes and provides targeted insights for improvement. By offering this standardized diagnostic framework, we aim to facilitate more robust development of audio-visual models. The dataset is released: https://github.com/gorjanradevski/dave

Unified Multimodal Models (UMMs) have shown remarkable progress in visual generation. Yet, existing benchmarks predominantly assess Crystallized Intelligence, which relies on recalling accumulated knowledge and learned schemas. This focus overlooks Generative Fluid Intelligence (GFI): the capacity to induce patterns, reason through constraints, and adapt to novel scenarios on the fly. To rigorously assess this capability, we introduce GENIUS (GEN Fluid Intelligence EvalUation Suite). We formalize GFI as a synthesis of three primitives. These include Inducing Implicit Patterns (e.g., inferring personalized visual preferences), Executing Ad-hoc Constraints (e.g., visualizing abstract metaphors), and Adapting to Contextual Knowledge (e.g., simulating counter-intuitive physics). Collectively, these primitives challenge models to solve problems grounded entirely in the immediate context. Our systematic evaluation of 12 representative models reveals significant performance deficits in these tasks. Crucially, our diagnostic analysis disentangles these failure modes. It demonstrates that deficits stem from limited context comprehension rather than insufficient intrinsic generative capability. To bridge this gap, we propose a training-free attention intervention strategy. Ultimately, GENIUS establishes a rigorous standard for GFI, guiding the field beyond knowledge utilization toward dynamic, general-purpose reasoning. Our dataset and code will be released at: https://github.com/arctanxarc/GENIUS{https://github.com/arctanxarc/GENIUS}.

Reliable depth estimation under real optical conditions remains a core challenge for camera vision in systems such as autonomous robotics and augmented reality. Despite recent progress in depth estimation and depth-of-field rendering, research remains constrained by the lack of large-scale, high-fidelity, real stereo DSLR datasets, limiting real-world generalization and evaluation of models trained on synthetic data as shown extensively in literature. We present the first high-resolution (5472times3648px) stereo DSLR dataset with 18000 images, systematically varying focal length and aperture across complex real scenes and capturing the optical realism and complexity of professional camera systems. For 9 scenes with varying scene complexity, lighting and background, images are captured with two identical camera assemblies at 10 focal lengths (28-70mm) and 5 apertures (f/2.8-f/22), spanning 50 optical configurations in 2000 images per scene. This full-range optics coverage enables controlled analysis of geometric and optical effects for monocular and stereo depth estimation, shallow depth-of-field rendering, deblurring, 3D scene reconstruction and novel view synthesis. Each focal configuration has a dedicated calibration image set, supporting evaluation of classical and learning based methods for intrinsic and extrinsic calibration. The dataset features challenging visual elements such as multi-scale optical illusions, reflective surfaces, mirrors, transparent glass walls, fine-grained details, and natural / artificial ambient light variations. This work attempts to bridge the realism gap between synthetic training data and real camera optics, and demonstrates challenges with the current state-of-the-art monocular, stereo depth and depth-of-field methods. We release the dataset, calibration files, and evaluation code to support reproducible research on real-world optical generalization.

Software engineering, whether performed by humans or by AI agents, requires reasoning about how software behaves. We call the internal model that supports such reasoning the software world model, and view current code-execution benchmarks as covering one well-studied slice of it -- control flow. In this paper, we take a step toward a broader evaluation by shifting the observable axis to execution resources: alongside test outcome and exception class, we predict peak memory, wall-clock time, and ranked profiler outputs at method and line granularity. We use SWE-bench Verified as the source of data to hold the test close to real-world software engineering tasks. All tested models, frontier ones included, show modest performance and brittle behaviour, suggesting a notable lack of understanding of how software is executed, as opposed to how its source code is written.

Implicit meanings are integral to human communication, making it essential for language models to be capable of identifying and interpreting them. Grice (1975) proposed a set of conversational maxims that guide cooperative dialogue, noting that speakers may deliberately violate these principles to express meanings beyond literal words, and that listeners, in turn, recognize such violations to draw pragmatic inferences. Building on Surian et al. (1996)'s study of children's sensitivity to violations of Gricean maxims, we introduce a novel benchmark to test whether language models pretrained on less than 10M and less than 100M tokens can distinguish maxim-adhering from maxim-violating utterances. We compare these BabyLMs across five maxims and situate their performance relative to children and a Large Language Model (LLM) pretrained on 3T tokens. We find that overall, models trained on less than 100M tokens outperform those trained on less than 10M, yet fall short of child-level and LLM competence. Our results suggest that modest data increases improve some aspects of pragmatic behavior, leading to finer-grained differentiation between pragmatic dimensions.

Current multimodal benchmarks often conflate reasoning with domain-specific knowledge, making it difficult to isolate and evaluate general reasoning abilities in non-expert settings. To address this, we introduce VisualPuzzles, a benchmark that targets visual reasoning while deliberately minimizing reliance on specialized knowledge. VisualPuzzles consists of diverse questions spanning five categories: algorithmic, analogical, deductive, inductive, and spatial reasoning. One major source of our questions is manually translated logical reasoning questions from the Chinese Civil Service Examination. Experiments show that VisualPuzzles requires significantly less intensive domain-specific knowledge and more complex reasoning compared to benchmarks like MMMU, enabling us to better evaluate genuine multimodal reasoning. Evaluations show that state-of-the-art multimodal large language models consistently lag behind human performance on VisualPuzzles, and that strong performance on knowledge-intensive benchmarks does not necessarily translate to success on reasoning-focused, knowledge-light tasks. Additionally, reasoning enhancements such as scaling up inference compute (with "thinking" modes) yield inconsistent gains across models and task types, and we observe no clear correlation between model size and performance. We also found that models exhibit different reasoning and answering patterns on VisualPuzzles compared to benchmarks with heavier emphasis on knowledge. VisualPuzzles offers a clearer lens through which to evaluate reasoning capabilities beyond factual recall and domain knowledge.

AI evaluation results are produced at scale but reported inconsistently across leaderboards, model cards, benchmark papers, and company blogs. The cost is interpretive: readers cannot reliably compare results across sources, identify what a report omits, or trace an aggregate claim to its underlying evidence. Recent efforts address isolated components but leave three gaps: they cover only narrow slices of the evaluation lifecycle and do not compose into a single interpretable record; they specify static representations that do not differentiate the questions different stakeholders bring to the same evidence; and they remain proposals on paper, lacking the extraction infrastructure required for adoption at scale. We present , an operational reporting layer that composes benchmark metadata, evaluation run data, and model metadata into a unified record. We (1) derive a reporting schema from a structured review of 52 papers and 10 stakeholder interviews, (2) implement four interpretive signals (reproducibility, documentation completeness, provenance and risk, and score comparability), rendered through reader modes calibrated to research and non-research audiences, and (3) deploy a monitoring tool that applies across 5,816 models, 635 benchmarks, and 101,843 results, surfacing systematic gaps in current reporting practice.

Customizing an LLM judge to a specific task or domain often involves optimizing its prompt across multiple evaluation criteria simultaneously. Textual gradient methods automate this for a single judge criterion, however they produce natural-language critiques, not numerical vectors. Thus, the conflict-resolution toolkit of multi-task learning (PCGrad, MGDA) doesn't apply to the multi-objective textual gradient setting. We test five decomposition modes of textual gradient optimizers by varying how much cross-task information the loss, gradient and optimizer LLMs share. In 6 of 10 configurations, we observe that optimization never improves over the initial prompt. Gradient specificity drops by 59% (from 9.0 to 3.7) when the gradient LLM processes multiple criteria jointly. Separately, we observe that naively combining per-task instructions into a single prompt degrades Spearman's rho by -5.3%. These results identify two separable failure modes: optimization-time gradient dilution and inference-time instruction interference, which together constrain the design space for multi-objective judge customization using textual feedback.

Generative video models are increasingly studied as implicit world models, yet evaluating whether they produce physically plausible 3D structure and motion remains challenging. Most existing video evaluation pipelines rely heavily on human judgment or learned graders, which can be subjective and weakly diagnostic for geometric failures. We introduce PDI-Bench (Perspective Distortion Index), a quantitative framework for auditing geometric coherence in generated videos. Given a generated clip, we obtain object-centric observations via segmentation and point tracking (e.g., SAM 2, MegaSaM, and CoTracker3), lift them to 3D world-space coordinates via monocular reconstruction, and compute a set of projective-geometry residuals capturing three failure dimensions: scale-depth alignment, 3D motion consistency, and 3D structural rigidity. To support systematic evaluation, we build PDI-Dataset, covering diverse scenarios designed to stress these geometric constraints. Across state-of-the-art video generators, PDI reveals consistent geometry-specific failure modes that are not captured by common perceptual metrics, and provides a diagnostic signal for progress toward physically grounded video generation and physical world model. Our code and dataset can be found at https://pdi-bench.github.io/.

Large Language Models (LLMs) have made notable progress in automated theorem proving, yet existing formal benchmarks remain limited in both mathematical coverage and difficulty. Most are concentrated in areas that are easier to formalize, such as algebra and elementary number theory, and provide limited coverage of subfields that require deeper reasoning, including mathematical analysis. To address this gap, we introduce MA-ProofBench, to the best of our knowledge, the first formal theorem-proving benchmark dedicated to Mathematical Analysis. The benchmark contains 200 formalized theorems covering 6 core topics and 27 subcategories, including measure and integration theory, complex analysis, and functional analysis. The problems are divided into two difficulty levels, an undergraduate level (Level I, 100 problems) and a Ph.D. qualifying level (Level II, 100 problems), to evaluate how well LLMs perform formal reasoning at different mathematical depths. Each problem is constructed through a human-led, LLM-assisted formalization pipeline followed by independent expert review, ensuring that the formal statements remain faithful to the original mathematics. We evaluate a range of recent general-purpose reasoning models and formal theorem provers on MA-ProofBench. However, most models perform poorly: even the best-performing model, GPT-5.5, achieves only 16% Pass@8 on Level I and 5% on Level II, while most models stay close to 0% on Level II. Further analysis identifies Mathlib hallucinations and incomplete proofs as the two dominant failure modes, while an evaluation on the natural-language version of the benchmark exposes a clear gap between informal and formal reasoning. MA-ProofBench is intended to serve as a reliable reference for tracking progress in formal mathematical reasoning in advanced domains.

Failures in multi-turn reasoning models are largely invisible to terminal-score evaluation. A model can lock onto an unsafe stance early in a long dialogue, yet its final-turn refusal rate may appear indistinguishable from a robustly aligned baseline. To expose these hidden temporal dynamics, we propose a trace-level diagnostic - the CoT-Output 2x2 safety matrix. This framework labels every turn along two independent axes (internal reasoning and visible output), yielding four operationally defined failure cells: robust alignment, alignment faking, overt jailbreak, and a distinct failure mode we term context-injection failure (where the CoT maintains safe reasoning, but the visible output produces harm, highlighting a multi-turn manifestation of reasoning unfaithfulness). We evaluate three distilled reasoning targets against a fixed attacker across five oversight conditions, collecting 6750 turn-level observations on the Information-Hazard scenario. Our analysis reveals two reproducible vulnerabilities: an oversight paradox where explicit monitoring cues paradoxically increase alignment-faking rates rather than suppress them, and a context-injection failure where models lock onto unsafe external outputs despite safe internal states. We release the full dataset of multi-turn dialogues and CoT traces to support follow-up trace-diagnostic research.

Communications satellite network (CSN), as an integral component of the next generation of communication systems, has the capability to offer services globally. Data transmission in this network primarily relies on two modes: inter-satellite communication and satellite-to-ground station communication. The latter directly impacts the successful reception of data by users. However, due to resource and task limitations, finding a satisfactory solution poses a significant challenge. The communication satellite-ground station network scheduling problem (CS-GSNSP) aims to optimize CSN effectiveness by devising a plan that maximizes link construction time while considering constraints associated with satellite operation modes. The large number of tasks and numerous constraints in the problem result in a time-consuming evaluation of fitness function values. To address this issue, we propose a fuzzy fitness evaluation method (FFEM) that employs fuzzy or real evaluation methods based on individual similarity degrees. Additionally, we introduce an evolutionary algorithm based on FFEM, called evolutionary algorithm based on FFEM (FFEEA), for iteratively searching high-quality network construction schemes. In FFEEA, an adaptive crossover approach is used for efficient population search. Finally, extensive experiments are conducted to demonstrate that our proposed fuzzy fitness evaluation method and other improvement strategies significantly enhance satellite network service time. The study introduces a novel approach to enhance the efficiency of solving combinatorial optimization problems, such as CS-GSNSP, by mitigating the complexity associated with fitness evaluation.

The rise of LLM-based agents has opened new frontiers in AI applications, yet evaluating these agents remains a complex and underdeveloped area. This survey provides an in-depth overview of the emerging field of LLM agent evaluation, introducing a two-dimensional taxonomy that organizes existing work along (1) evaluation objectives -- what to evaluate, such as agent behavior, capabilities, reliability, and safety -- and (2) evaluation process -- how to evaluate, including interaction modes, datasets and benchmarks, metric computation methods, and tooling. In addition to taxonomy, we highlight enterprise-specific challenges, such as role-based access to data, the need for reliability guarantees, dynamic and long-horizon interactions, and compliance, which are often overlooked in current research. We also identify future research directions, including holistic, more realistic, and scalable evaluation. This work aims to bring clarity to the fragmented landscape of agent evaluation and provide a framework for systematic assessment, enabling researchers and practitioners to evaluate LLM agents for real-world deployment.

Evaluations of model editing currently only use the `next few token' completions after a prompt. As a result, the impact of these methods on longer natural language generation is largely unknown. We introduce long-form evaluation of model editing (LEME) a novel evaluation protocol that measures the efficacy and impact of model editing in long-form generative settings. Our protocol consists of a machine-rated survey and a classifier which correlates well with human ratings. Importantly, we find that our protocol has very little relationship with previous short-form metrics (despite being designed to extend efficacy, generalization, locality, and portability into a long-form setting), indicating that our method introduces a novel set of dimensions for understanding model editing methods. Using this protocol, we benchmark a number of model editing techniques and present several findings including that, while some methods (ROME and MEMIT) perform well in making consistent edits within a limited scope, they suffer much more from factual drift than other methods. Finally, we present a qualitative analysis that illustrates common failure modes in long-form generative settings including internal consistency, lexical cohesion, and locality issues.

As Large Language Models (LLMs) evolve into autonomous agents, the need for unified evaluation infrastructure becomes critical. However, current evaluation pipelines remain highly fragmented and tightly coupled, hindering reproducibility and causing redundant engineering. To address this, we introduce AgentCompass, an open-source, lightweight, and extensible infrastructure for evaluating LLM-based agents. AgentCompass organizes the evaluation process around three independent components, namely Benchmark, Harness, and Environment, thereby enabling flexible configurations without requiring the reimplementation of complex execution logic. Furthermore, it features a fault-tolerant asynchronous runtime and comprehensive trajectory analysis tools to transparently diagnose nuanced failure modes like reward-hacking. Natively supporting over 20 benchmarks across five capability dimensions, AgentCompass provides the community with a scalable and reproducible infrastructure for advancing agent research.

Large language models (LLMs) are increasingly deployed as customer-facing agents, yet evaluating their reliability remains challenging due to stochastic, multi-turn interactions. Current evaluation protocols rely on linear Monte Carlo rollouts of complete agent-user conversations to estimate success. However, this approach is computationally inefficient, repeatedly regenerating identical early prefixes, and often fails to uncover deep failure modes that arise from rare user behaviors. We introduce DIVERT (Diversity-Induced Evaluation via Branching of Trajectories), an efficient, snapshot-based, coverage-guided user simulation framework for systematic exploration of agent-user interactions. DIVERT captures the full agent-environment state at critical decision points and resumes execution from these snapshots, enabling reuse of shared conversation prefixes and reducing redundant computation. From each junction, the framework branches using targeted, diversity-inducing user responses, allowing directed exploration of alternative interaction paths. By focusing evaluation on semantically diverse and underexplored trajectories, DIVERT improves both efficiency and coverage. Empirical results show that it discovers more failures per token compared to standard linear rollout protocols, while expanding the set of tasks on which failures are identified.

Multi-provider LLM gateways reverse proxies that route, load-balance, and rate-limit requests across foundation-model APIs have become critical production infrastructure. Yet the failure modes specific to this architectural layer remain undocumented, scattered across issue trackers and post-mortems with no unifying framework. We introduce , a two-axis taxonomy that classifies failures by their origin layer (Network/Transport, Streaming/Protocol, State/Session, Model~Behavior, Governance/Cost) and their detectability (Loud vs.\ Silent). We populate this taxonomy with five verified catalog entries sourced from public bug reports and first-hand stress testing, each accompanied by a mechanistic root-cause analysis. Three entries include standalone reproduction scripts. Our principal finding is that the most operationally severe failures are silent: they return HTTP~200, pass every standard health check, and corrupt application state in ways that require semantic-level observability to detect. Two such silent failures a concurrency race condition causing history loss and a streaming index collision corrupting tool-call payloads were discovered first-hand during evaluation campaigns.

Evaluating text-to-SQL systems remains largely fragile: correctness is typically judged by executing predicted and gold SQL queries on a single static database, even though the same queries may behave differently under alternative database instances. This raises a broader language modeling question: Can large language models synthesize semantically meaningful, schema-consistent relational data directly from a natural language question? If so, such generation can serve as a controlled mechanism for stress-testing text-to-SQL systems beyond fixed benchmark databases. We introduce SynSQL, a framework that synthesizes test databases conditioned on question-schema alignment rather than gold SQL queries. SynSQL decomposes the task into three stages: (1) schema selection, (2) question-guided data synthesis, and (3) constraint-aware critique with iterative refinement, framing database construction as structured generation under semantic and relational constraints. Across ten text-to-SQL models on Spider, BIRD, and Spider 2.0, SynSQL-generated databases reveal performance drops of 3-14% compared to static evaluation, exposing errors masked by benchmark artifacts. We further analyze generation quality, constraint adherence, and failure modes, highlighting both the promise and limitations of LLMs in structured data synthesis. Our findings position synthetic database generation as a new lens for studying LLM reasoning, controllability, and robustness in structured environments.

Rubric-based evaluation has become a prevailing paradigm for evaluating instruction following in large language models (LLMs). Despite its widespread use, the reliability of these rubric-level evaluations remains unclear, calling for meta-evaluation. However, prior meta-evaluation efforts largely focus on the response level, failing to assess the fine-grained judgment accuracy that rubric-based evaluation relies on. To bridge this gap, we introduce RubricEval. Our benchmark features: (1) the first rubric-level meta-evaluation benchmark for instruction following, (2) diverse instructions and responses spanning multiple categories and model sources, and (3) a substantial set of 3,486 quality-controlled instances, along with Easy/Hard subsets that better differentiates judge performance. Our experiments reveal that rubric-level judging remains far from solved: even GPT-4o, a widely adopted judge in instruction-following benchmarks, achieves only 55.97% on Hard subset. Considering evaluation paradigm, rubric-level evaluation outperforms checklist-level, explicit reasoning improves accuracy, and both together reduce inter-judge variance. Through our established rubric taxonomy, we further identify common failure modes and offer actionable insights for reliable instruction-following evaluation.

The LLM-as-a-Judge paradigm promises scalable rubric-based evaluation, yet aligning frozen black-box models with human standards remains a challenge due to inherent generation stochasticity. We reframe judge alignment as a criteria transfer problem and isolate three recurrent failure modes: rubric instability caused by prompt sensitivity, unverifiable reasoning that lacks auditable evidence, and scale misalignment with human grading boundaries. To address these issues, we introduce RULERS (Rubric Unification, Locking, and Evidence-anchored Robust Scoring), a compiler-executor framework that transforms natural language rubrics into executable specifications. RULERS operates by compiling criteria into versioned immutable bundles, enforcing structured decoding with deterministic evidence verification, and applying lightweight Wasserstein-based post-hoc calibration, all without updating model parameters. Extensive experiments on essay and summarization benchmarks demonstrate that RULERS significantly outperforms representative baselines in human agreement, maintains strong stability against adversarial rubric perturbations, and enables smaller models to rival larger proprietary judges. Overall, our results suggest that reliable LLM judging requires executable rubrics, verifiable evidence, and calibrated scales rather than prompt phrasing alone. Code is available at https://github.com/LabRAI/Rulers.git.

Multimodal Large Language Models (MLLMs) are increasingly used for content moderation, yet their robustness in short-form video contexts remains underexplored. Current safety evaluations often rely on unimodal attacks, failing to address combined attack vulnerabilities. In this paper, we introduce a comprehensive framework for evaluating the tri-modal safety of MLLMs. First, we present the Short-Video Multimodal Adversarial (SVMA) dataset, comprising diverse short-form videos with human-guided synthetic adversarial attacks. Second, we propose ChimeraBreak, a novel tri-modal attack strategy that simultaneously challenges visual, auditory, and semantic reasoning pathways. Extensive experiments on state-of-the-art MLLMs reveal significant vulnerabilities with high Attack Success Rates (ASR). Our findings uncover distinct failure modes, showing model biases toward misclassifying benign or policy-violating content. We assess results using LLM-as-a-judge, demonstrating attack reasoning efficacy. Our dataset and findings provide crucial insights for developing more robust and safe MLLMs.

Large language models (LLMs) have achieved remarkable capabilities but remain vulnerable to adversarial prompts known as jailbreaks, which can bypass safety alignment and elicit harmful outputs. Despite growing efforts in LLM safety research, existing evaluations are often fragmented, focused on isolated attack or defense techniques, and lack systematic, reproducible analysis. In this work, we introduce PandaGuard, a unified and modular framework that models LLM jailbreak safety as a multi-agent system comprising attackers, defenders, and judges. Our framework implements 19 attack methods and 12 defense mechanisms, along with multiple judgment strategies, all within a flexible plugin architecture supporting diverse LLM interfaces, multiple interaction modes, and configuration-driven experimentation that enhances reproducibility and practical deployment. Built on this framework, we develop PandaBench, a comprehensive benchmark that evaluates the interactions between these attack/defense methods across 49 LLMs and various judgment approaches, requiring over 3 billion tokens to execute. Our extensive evaluation reveals key insights into model vulnerabilities, defense cost-performance trade-offs, and judge consistency. We find that no single defense is optimal across all dimensions and that judge disagreement introduces nontrivial variance in safety assessments. We release the code, configurations, and evaluation results to support transparent and reproducible research in LLM safety.

Diversity is an important criterion for many areas of machine learning (ML), including generative modeling and dataset curation. Yet little work has gone into understanding, formalizing, and measuring diversity in ML. In this paper, we address the diversity evaluation problem by proposing the Vendi Score, which connects and extends ideas from ecology and quantum statistical mechanics to ML. The Vendi Score is defined as the exponential of the Shannon entropy of the eigenvalues of a similarity matrix. This matrix is induced by a user-defined similarity function applied to the sample to be evaluated for diversity. In taking a similarity function as input, the Vendi Score enables its user to specify any desired form of diversity. Importantly, unlike many existing metrics in ML, the Vendi Score doesn't require a reference dataset or distribution over samples or labels, it is therefore general and applicable to any generative model, decoding algorithm, and dataset from any domain where similarity can be defined. We showcased the Vendi Score on molecular generative modeling, a domain where diversity plays an important role in enabling the discovery of novel molecules. We found that the Vendi Score addresses shortcomings of the current diversity metric of choice in that domain. We also applied the Vendi Score to generative models of images and decoding algorithms of text and found it confirms known results about diversity in those domains. Furthermore, we used the Vendi Score to measure mode collapse, a known limitation of generative adversarial networks (GANs). In particular, the Vendi Score revealed that even GANs that capture all the modes of a labeled dataset can be less diverse than the original dataset. Finally, the interpretability of the Vendi Score allowed us to diagnose several benchmark ML datasets for diversity, opening the door for diversity-informed data augmentation.

Recent studies demonstrate that tool-calling capability enables large language models (LLMs) to interact with external environments for long-horizon financial tasks. While existing benchmarks have begun evaluating financial tool calling, they focus on limited scenarios and rely on call-level metrics that fail to capture trajectory-level reasoning quality. To address this gap, we introduce FinTrace, a benchmark comprising 800 expert-annotated trajectories spanning 34 real-world financial task categories across multiple difficulty levels. FinTrace employs a rubric-based evaluation protocol with nine metrics organized along four axes -- action correctness, execution efficiency, process quality, and output quality -- enabling fine-grained assessment of LLM tool-calling behavior. Our evaluation of 13 LLMs reveals that while frontier models achieve strong tool selection, all models struggle with information utilization and final answer quality, exposing a critical gap between invoking the right tools and reasoning effectively over their outputs. To move beyond diagnosis, we construct FinTrace-Training, the first trajectory-level preference dataset for financial tool-calling, containing 8,196 curated trajectories with tool-augmented contexts and preference pairs. We fine-tune Qwen-3.5-9B using supervised fine-tuning followed by direct preference optimization (DPO) and show that training on FinTrace-Training consistently improves intermediate reasoning metrics, with DPO more effectively suppressing failure modes. However, end-to-end answer quality remains a bottleneck, indicating that trajectory-level improvements do not yet fully propagate to final output quality.

Recent years have seen impressive advances in text-to-image generation, with image generative or unified models producing high-quality images from text. Yet these models still struggle with fine-grained color controllability, often failing to accurately match colors specified in text prompts. While existing benchmarks evaluate compositional reasoning and prompt adherence, none systematically assess color precision. Color is fundamental to human visual perception and communication, critical for applications from art to design workflows requiring brand consistency. However, current benchmarks either neglect color or rely on coarse assessments, missing key capabilities such as interpreting RGB values or aligning with human expectations. To this end, we propose GenColorBench, the first comprehensive benchmark for text-to-image color generation, grounded in color systems like ISCC-NBS and CSS3/X11, including numerical colors which are absent elsewhere. With 44K color-focused prompts covering 400+ colors, it reveals models' true capabilities via perceptual and automated assessments. Evaluations of popular text-to-image models using GenColorBench show performance variations, highlighting which color conventions models understand best and identifying failure modes. Our GenColorBench assessments will guide improvements in precise color generation. The benchmark will be made public upon acceptance.

Text-to-image (T2I) generative models, such as Stable Diffusion and DALL-E, have shown remarkable proficiency in producing high-quality, realistic, and natural images from textual descriptions. However, these models sometimes fail to accurately capture all the details specified in the input prompts, particularly concerning entities, attributes, and spatial relationships. This issue becomes more pronounced when the prompt contains novel or complex compositions, leading to what are known as compositional generation failure modes. Recently, a new open-source diffusion-based T2I model, FLUX, has been introduced, demonstrating strong performance in high-quality image generation. Additionally, autoregressive T2I models like LlamaGen have claimed competitive visual quality performance compared to diffusion-based models. In this study, we evaluate the compositional generation capabilities of these newly introduced models against established models using the T2I-CompBench benchmark. Our findings reveal that LlamaGen, as a vanilla autoregressive model, is not yet on par with state-of-the-art diffusion models for compositional generation tasks under the same criteria, such as model size and inference time. On the other hand, the open-source diffusion-based model FLUX exhibits compositional generation capabilities comparable to the state-of-the-art closed-source model DALL-E3.

Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes. This paper aggregates the largest coordinated deep-dive of one MCP-based industrial-agent benchmark to date: fourteen parallel implementation studies covering new asset classes (including a multi-modal visual extension), alternative orchestrations, retrieval strategies, reasoning modes, infrastructure optimizations, and evaluation-methodology probes. Consolidating those studies with seven prior agent benchmarks, we argue that aggregate-score leaderboards systematically underspecify deployed-agent evaluation. Rankings derived from aggregate scores do not transfer to out-of-distribution settings; recent public-to-hidden competition retrospectives provide direct empirical evidence of this rank instability. We propose ranking configurations by predictive validity, the correlation between in-sample and out-of-sample rank, rather than in-sample mean, and report a twelve-tier measurement apparatus that exposes the deployment-relevant dimensions HELM and its agent-era successors collapse. The position is operationalized through three falsifiable out-of-distribution criteria with explicit thresholds; existing evidence partly supports it but is too thin to confirm. We close with a pre-registered pilot design and a field-level vision for what the next generation of agentic benchmarks should report.

Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities, yet their proficiency in understanding and reasoning over multiple images remains largely unexplored. While existing benchmarks have initiated the evaluation of multi-image models, a comprehensive analysis of their core weaknesses and their causes is still lacking. In this work, we introduce MIMIC (Multi-Image Model Insights and Challenges), a new benchmark designed to rigorously evaluate the multi-image capabilities of LVLMs. Using MIMIC, we conduct a series of diagnostic experiments that reveal pervasive issues: LVLMs often fail to aggregate information across images and struggle to track or attend to multiple concepts simultaneously. To address these failures, we propose two novel complementary remedies. On the data side, we present a procedural data-generation strategy that composes single-image annotations into rich, targeted multi-image training examples. On the optimization side, we analyze layer-wise attention patterns and derive an attention-masking scheme tailored for multi-image inputs. Experiments substantially improved cross-image aggregation, while also enhancing performance on existing multi-image benchmarks, outperforming prior state of the art across tasks. Data and code will be made available at https://github.com/anurag-198/MIMIC.

Instruction-based image editing has advanced rapidly, yet reliable and interpretable evaluation remains a bottleneck. Current protocols either (i) depend on paired reference images -- resulting in limited coverage and inheriting biases from prior generative models -- or (ii) rely solely on zero-shot vision-language models (VLMs), whose prompt-based assessments of instruction following, content consistency, and visual quality are often imprecise. To address this, we introduce EdiVal-Agent, an automated, scalable, and fine-grained evaluation framework for multi-turn instruction-based editing from an object-centric perspective, supported by a suite of expert tools. Given an image, EdiVal-Agent first decomposes it into semantically meaningful objects, then synthesizes diverse, context-aware editing instructions. For evaluation, it integrates VLMs with open-vocabulary object detectors to assess instruction following, uses semantic-level feature extractors to evaluate content consistency, and leverages human preference models to judge visual quality. We show that combining VLMs with object detectors yields stronger agreement with human judgments in instruction-following evaluation compared to using VLMs alone and CLIP-based metrics. Furthermore, the pipeline's modular design allows future tools to be seamlessly integrated, enhancing evaluation accuracy over time. Instantiating this pipeline, we build EdiVal-Bench, a multi-turn editing benchmark covering 9 instruction types and 11 state-of-the-art editing models spanning autoregressive (AR) (including Nano Banana, GPT-Image-1), flow-matching, and diffusion paradigms. We demonstrate that EdiVal-Agent can be used to identify existing failure modes, thereby informing the development of the next generation of editing models. Project page: https://tianyucodings.github.io/EdiVAL-page/.

Current robotic evaluation is still largely dominated by binary success rates, which collapse rich execution processes into a single outcome and obscure critical qualities such as progress, efficiency, and stability. To address this limitation, we propose PRM-as-a-Judge, a dense evaluation paradigm that leverages Process Reward Models (PRMs) to audit policy execution directly from trajectory videos by estimating task progress from observation sequences. Central to this paradigm is the OPD (Outcome-Process-Diagnosis) metric system, which explicitly formalizes execution quality via a task-aligned progress potential. We characterize dense robotic evaluation through two axiomatic properties: macro-consistency, which requires additive and path-consistent aggregation, and micro-resolution, which requires sensitivity to fine-grained physical evolution. Under this formulation, potential-based PRM judges provide a natural instantiation of dense evaluation, with macro-consistency following directly from the induced scalar potential. We empirically validate the micro-resolution property using RoboPulse, a diagnostic benchmark specifically designed for probing micro-scale progress discrimination, where several trajectory-trained PRM judges outperform discriminative similarity-based methods and general-purpose foundation-model judges. Finally, leveraging PRM-as-a-Judge and the OPD metric system, we conduct a structured audit of mainstream policy paradigms across long-horizon tasks, revealing behavioral signatures and failure modes that are invisible to outcome-only metrics.

Evaluating the quality of multi-turn chatbot interactions remains challenging, as most existing methods assess interactions at the turn level without addressing whether a user's overarching goal was fulfilled. A ``goal'' here refers to an information need or task, such as asking for policy information or applying for leave. We propose a comprehensive framework for goal-oriented evaluation of multi-agent systems (MAS), introducing the Goal Success Rate (GSR) to measure the percentage of fulfilled goals, and a Root Cause of Failure (RCOF) taxonomy to identify reasons for failure in multi-agent chatbots. Our method segments conversations by user goals and evaluates success using all relevant turns. We present a model-based evaluation system combining teacher LLMs, where domain experts define goals, set quality standards serving as a guidance for the LLMs. The LLMs use ``thinking tokens'' to produce interpretable rationales, enabling explainable, data-efficient evaluations. In an enterprise setting, we apply our framework to evaluate AIDA, a zero-to-one employee conversational agent system built as a ground-up multi-agent conversational agent, and observe GSR improvement from 63\% to 79\% over six months since its inception. Our framework is generic and offers actionable insights through a detailed defect taxonomy based on analysis of failure points in multi-agent chatbots, diagnosing overall success, identifying key failure modes, and informing system improvements.

Reliable benchmarking is critical for developing machine learning models for tandem mass spectrometry (MS/MS) based molecule discovery. Subtle issues in experimental design and model evaluation procedures can degrade the trustworthiness of such benchmarks and lead to erroneous conclusions. We conduct a thorough review of model evaluation issues in the recent MS/MS machine learning literature, using the standard MassSpecGym benchmark suite as a case study to illustrate the impact of these issues. We find evaluation issues in at least 17 of 26 papers reporting MassSpecGym benchmark results in the first year of its adoption. We isolate three classes of failures: (i) data leakage, (ii) shortcut learning, and (iii) implementation bugs and metric divergence. Through extensive experimentation and code replication, we quantify the impact of these issues and show how they corrupt the evaluation standards MassSpecGym was designed to enforce. We distill our findings into recommendations generalizable to MS/MS challenges, benchmarks, and custom evaluation setups. We also release MassSpecGym v1.5, an implementation of our recommendations in the MassSpecGym benchmarking suite which addresses the failure modes identified in this audit. MassSpecGym v1.5 is publicly available at https://github.com/pluskal-lab/MassSpecGym.

Large reasoning models (e.g., R1, o3) have demonstrated remarkable mathematical problem-solving abilities. However, the high reported accuracy of these advanced models on popular datasets, reliance on purely numerical evaluation and potential benchmark leakage, often masks their true reasoning shortcomings. To address this, we propose leveraging the inherent rigor and methodological complexity of mathematical proofs as a diagnostic tool to expose these hidden failures. Specifically, we introduce the RFMDataset (Reveal Failure Modes), a collection of 200 diverse mathematical proof problems, and thoroughly evaluate advanced models' performance on it. Our in-depth analysis of their failures uncovers 10 fine-grained error types, which shows fundamental limitations in current large reasoning models: 1) large reasoning models grapple profoundly with mathematical proofs, with some generating entirely correct proofs for less than 20% of problems and failing even on basic ones; 2) models exhibit a diverse spectrum of reasoning failures, prominently demonstrating the lack of guarantees for the correctness and rigor of single-step reasoning; and 3) models show hallucination and incompleteness during the reasoning process. Our findings reveal that models' self-reflection is insufficient to resolve the current logical dilemmas, necessitating formalized and fine-grained logical training.

AI agents are increasingly deployed in production, yet their security evaluations remain bottlenecked by manual red-teaming or static benchmarks that fail to model adaptive, multi-turn adversaries. We propose NAAMSE, an evolutionary framework that reframes agent security evaluation as a feedback-driven optimization problem. Our system employs a single autonomous agent that orchestrates a lifecycle of genetic prompt mutation, hierarchical corpus exploration, and asymmetric behavioral scoring. By using model responses as a fitness signal, the framework iteratively compounds effective attack strategies while simultaneously ensuring "benign-use correctness", preventing the degenerate security of blanket refusal. Our experiments on Gemini 2.5 Flash demonstrate that evolutionary mutation systematically amplifies vulnerabilities missed by one-shot methods, with controlled ablations revealing that the synergy between exploration and targeted mutation uncovers high-severity failure modes. We show that this adaptive approach provides a more realistic and scalable assessment of agent robustness in the face of evolving threats. The code for NAAMSE is open source and available at https://github.com/HASHIRU-AI/NAAMSE.

Deep Research Agents (DRAs) aim to solve complex, long-horizon research tasks involving planning, retrieval, multimodal understanding, and report generation, yet their evaluation remains challenging due to dynamic web environments and ambiguous task definitions. We propose DR^{3}-Eval, a realistic and reproducible benchmark for evaluating deep research agents on multimodal, multi-file report generation. DR^{3}-Eval is constructed from authentic user-provided materials and paired with a per-task static research sandbox corpus that simulates open-web complexity while remaining fully verifiable, containing supportive documents, distractors, and noise. Moreover, we introduce a multi-dimensional evaluation framework measuring Information Recall, Factual Accuracy, Citation Coverage, Instruction Following, and Depth Quality, and validate its alignment with human judgments. Experiments with our developed multi-agent system DR^{3}-Agent based on multiple state-of-the-art language models demonstrate that DR^{3}-Eval is highly challenging and reveals critical failure modes in retrieval robustness and hallucination control. Our code and data are publicly available.

We introduce FailureSensorIQ, a novel Multi-Choice Question-Answering (MCQA) benchmarking system designed to assess the ability of Large Language Models (LLMs) to reason and understand complex, domain-specific scenarios in Industry 4.0. Unlike traditional QA benchmarks, our system focuses on multiple aspects of reasoning through failure modes, sensor data, and the relationships between them across various industrial assets. Through this work, we envision a paradigm shift where modeling decisions are not only data-driven using statistical tools like correlation analysis and significance tests, but also domain-driven by specialized LLMs which can reason about the key contributors and useful patterns that can be captured with feature engineering. We evaluate the Industrial knowledge of over a dozen LLMs-including GPT-4, Llama, and Mistral-on FailureSensorIQ from different lens using Perturbation-Uncertainty-Complexity analysis, Expert Evaluation study, Asset-Specific Knowledge Gap analysis, ReAct agent using external knowledge-bases. Even though closed-source models with strong reasoning capabilities approach expert-level performance, the comprehensive benchmark reveals a significant drop in performance that is fragile to perturbations, distractions, and inherent knowledge gaps in the models. We also provide a real-world case study of how LLMs can drive the modeling decisions on 3 different failure prediction datasets related to various assets. We release: (a) expert-curated MCQA for various industrial assets, (b) FailureSensorIQ benchmark and Hugging Face leaderboard based on MCQA built from non-textual data found in ISO documents, and (c) LLMFeatureSelector, an LLM-based feature selection scikit-learn pipeline. The software is available at https://github.com/IBM/FailureSensorIQ.

Quantitative evaluation has increased dramatically among recent video inpainting work, but the video and mask content used to gauge performance has received relatively little attention. Although attributes such as camera and background scene motion inherently change the difficulty of the task and affect methods differently, existing evaluation schemes fail to control for them, thereby providing minimal insight into inpainting failure modes. To address this gap, we propose the Diagnostic Evaluation of Video Inpainting on Landscapes (DEVIL) benchmark, which consists of two contributions: (i) a novel dataset of videos and masks labeled according to several key inpainting failure modes, and (ii) an evaluation scheme that samples slices of the dataset characterized by a fixed content attribute, and scores performance on each slice according to reconstruction, realism, and temporal consistency quality. By revealing systematic changes in performance induced by particular characteristics of the input content, our challenging benchmark enables more insightful analysis into video inpainting methods and serves as an invaluable diagnostic tool for the field. Our code and data are available at https://github.com/MichiganCOG/devil .

Advances in diffusion, autoregressive, and hybrid models have enabled high-quality image synthesis for tasks such as text-to-image, editing, and reference-guided composition. Yet, existing benchmarks remain limited, either focus on isolated tasks, cover only narrow domains, or provide opaque scores without explaining failure modes. We introduce ImagenWorld, a benchmark of 3.6K condition sets spanning six core tasks (generation and editing, with single or multiple references) and six topical domains (artworks, photorealistic images, information graphics, textual graphics, computer graphics, and screenshots). The benchmark is supported by 20K fine-grained human annotations and an explainable evaluation schema that tags localized object-level and segment-level errors, complementing automated VLM-based metrics. Our large-scale evaluation of 14 models yields several insights: (1) models typically struggle more in editing tasks than in generation tasks, especially in local edits. (2) models excel in artistic and photorealistic settings but struggle with symbolic and text-heavy domains such as screenshots and information graphics. (3) closed-source systems lead overall, while targeted data curation (e.g., Qwen-Image) narrows the gap in text-heavy cases. (4) modern VLM-based metrics achieve Kendall accuracies up to 0.79, approximating human ranking, but fall short of fine-grained, explainable error attribution. ImagenWorld provides both a rigorous benchmark and a diagnostic tool to advance robust image generation.

We introduce RoboEval, a structured evaluation framework and benchmark for robotic manipulation that augments binary success with principled behavioral and outcome metrics. Existing evaluations often collapse performance into outcome counts, masking differences in execution quality and obscuring failure structure. RoboEval provides eight bimanual tasks with systematically controlled variations, more than three thousand expert demonstrations, and a modular simulation platform for reproducible experimentation. All tasks are instrumented with standardized metrics that quantify efficiency, coordination, and safety/stability, as well as outcome measures that trace stagewise progress and localize failure modes. Through extensive experiments with state-of-the-art visuomotor policies, we validate these metrics by analyzing their stability under variation, discriminative power across policies with similar success rates, and correlation with task success. Project Page: https://robo-eval.github.io

The robot learning community has made great strides in recent years, proposing new architectures and showcasing impressive new capabilities; however, the dominant metric used in the literature, especially for physical experiments, is "success rate", i.e. the percentage of runs that were successful. Furthermore, it is common for papers to report this number with little to no information regarding the number of runs, the initial conditions, and the success criteria, little to no narrative description of the behaviors and failures observed, and little to no statistical analysis of the findings. In this paper we argue that to move the field forward, researchers should provide a nuanced evaluation of their methods, especially when evaluating and comparing learned policies on physical robots. To do so, we propose best practices for future evaluations: explicitly reporting the experimental conditions, evaluating several metrics designed to complement success rate, conducting statistical analysis, and adding a qualitative description of failures modes. We illustrate these through an evaluation on physical robots of several learned policies for manipulation tasks.

Most vehicle predictive maintenance systems rely exclusively on internal diagnostic signals and are validated on deterministic synthetic data, limiting the credibility of reported metrics. This paper presents a simulation-validated proof-of-concept framework for V2X-augmented predictive maintenance, integrating on-board sensor streams with external contextual signals -- road quality, weather, traffic density, and driver behaviour -- acquired via V2X communication and third-party APIs, with inference at the vehicle edge. Field validation on instrumented vehicles is identified as the required next step. Three experiments address common shortcomings of prior work. A feature group ablation study shows that V2X contextual features contribute a 2.6-point F1 gain, with full context removal reducing macro F1 from 0.855 to 0.807. On the AI4I 2020 real-world industrial failure dataset (10,000 samples, five failure modes), LightGBM achieves AUC-ROC of 0.973 under 5-fold stratified CV with SMOTE confined to training folds. A noise sensitivity analysis shows macro F1 remains above 0.88 under low noise and degrades to 0.74 under very high noise. SHAP analysis confirms that V2X and engineered interaction features rank among the top 15 predictors. Edge inference is estimated to reduce latency from 3.5s to under 1.0s versus cloud-only processing.

The development of Automatic Speech Recognition (ASR) systems for low-resource African languages remains challenging due to limited transcribed speech data. While recent advances in large multilingual models like OpenAI's Whisper offer promising pathways for low-resource ASR development, critical questions persist regarding practical deployment requirements. This paper addresses two fundamental concerns for practitioners: determining the minimum data volumes needed for viable performance and characterizing the primary failure modes that emerge in production systems. We evaluate Whisper's performance through comprehensive experiments on two Bantu languages: systematic data scaling analysis on Kinyarwanda using training sets from 1 to 1,400 hours, and detailed error characterization on Kikuyu using 270 hours of training data. Our scaling experiments demonstrate that practical ASR performance (WER < 13\%) becomes achievable with as little as 50 hours of training data, with substantial improvements continuing through 200 hours (WER < 10\%). Complementing these volume-focused findings, our error analysis reveals that data quality issues, particularly noisy ground truth transcriptions, account for 38.6\% of high-error cases, indicating that careful data curation is as critical as data volume for robust system performance. These results provide actionable benchmarks and deployment guidance for teams developing ASR systems across similar low-resource language contexts. We release accompanying and models see https://github.com/SunbirdAI/kinyarwanda-whisper-eval

Despite the growing use of large language models (LLMs) in clinical dialogue systems, existing evaluations focus on task completion or fluency, offering little insight into the behavioral and risk management requirements essential for safety-critical systems. This paper presents MATRIX (Multi-Agent simulaTion fRamework for safe Interactions and conteXtual clinical conversational evaluation), a structured, extensible framework for safety-oriented evaluation of clinical dialogue agents. MATRIX integrates three components: (1) a safety-aligned taxonomy of clinical scenarios, expected system behaviors and failure modes derived through structured safety engineering methods; (2) BehvJudge, an LLM-based evaluator for detecting safety-relevant dialogue failures, validated against expert clinician annotations; and (3) PatBot, a simulated patient agent capable of producing diverse, scenario-conditioned responses, evaluated for realism and behavioral fidelity with human factors expertise, and a patient-preference study. Across three experiments, we show that MATRIX enables systematic, scalable safety evaluation. BehvJudge with Gemini 2.5-Pro achieves expert-level hazard detection (F1 0.96, sensitivity 0.999), outperforming clinicians in a blinded assessment of 240 dialogues. We also conducted one of the first realism analyses of LLM-based patient simulation, showing that PatBot reliably simulates realistic patient behavior in quantitative and qualitative evaluations. Using MATRIX, we demonstrate its effectiveness in benchmarking five LLM agents across 2,100 simulated dialogues spanning 14 hazard scenarios and 10 clinical domains. MATRIX is the first framework to unify structured safety engineering with scalable, validated conversational AI evaluation, enabling regulator-aligned safety auditing. We release all evaluation tools, prompts, structured scenarios, and datasets.

We adapt Microsoft's QuantumKatas -- a well-established quantum computing curriculum -- from Q# to Qiskit, the most widely-adopted quantum computing framework, and package it with an evaluation framework for systematic LLM assessment. The resulting benchmark comprises 350 tasks across 26 categories, spanning fundamental gates through advanced algorithms (Grover's, Simon's, Deutsch-Jozsa), error correction, key distribution, and quantum games. Each task includes a natural language prompt, canonical solution, and deterministic test verification via classical circuit simulation. By building on the QuantumKatas' proven pedagogical design rather than creating tasks from scratch, we inherit a principled difficulty progression and comprehensive concept coverage while contributing the framework adaptation, evaluation infrastructure, and empirical analysis. We evaluate 16 LLMs across 7 prompting configurations -- a total of 39,200 model runs -- to demonstrate the benchmark's utility. Three key findings emerge: (1) the benchmark effectively differentiates model capabilities, with best-configuration pass rates ranging from 32.3% to 83.1% and a 26.1 pp average gap between frontier and open-source models; (2) models perform well at implementing known algorithms (SimonsAlgorithm 82.1%, BasicGates 81.6%) but struggle with problem encoding (SolveSATWithGrover 34.4%, DistinguishUnitaries 40.0%); and (3) chain-of-thought prompting shows a modestly bimodal effect -- it is the best strategy for three models (two of them explicitly reasoning-tuned per vendor documentation) but degrades performance for the rest, leaving it mid-pack in aggregate (56.3% mean) behind few-shot-5 (57.8%). We release the benchmark, evaluation framework, and baseline results to support research on LLM capabilities in quantum computing.

Out-of-distribution (OOD) detection is the problem of identifying inputs which are unrelated to the in-distribution task. The OOD detection performance when the in-distribution (ID) is ImageNet-1K is commonly being tested on a small range of test OOD datasets. We find that most of the currently used test OOD datasets, including datasets from the open set recognition (OSR) literature, have severe issues: In some cases more than 50% of the dataset contains objects belonging to one of the ID classes. These erroneous samples heavily distort the evaluation of OOD detectors. As a solution, we introduce with NINCO a novel test OOD dataset, each sample checked to be ID free, which with its fine-grained range of OOD classes allows for a detailed analysis of an OOD detector's strengths and failure modes, particularly when paired with a number of synthetic "OOD unit-tests". We provide detailed evaluations across a large set of architectures and OOD detection methods on NINCO and the unit-tests, revealing new insights about model weaknesses and the effects of pretraining on OOD detection performance. We provide code and data at https://github.com/j-cb/NINCO.

Novel neural architectures, training strategies, and the availability of large-scale corpora haven been the driving force behind recent progress in abstractive text summarization. However, due to the black-box nature of neural models, uninformative evaluation metrics, and scarce tooling for model and data analysis, the true performance and failure modes of summarization models remain largely unknown. To address this limitation, we introduce SummVis, an open-source tool for visualizing abstractive summaries that enables fine-grained analysis of the models, data, and evaluation metrics associated with text summarization. Through its lexical and semantic visualizations, the tools offers an easy entry point for in-depth model prediction exploration across important dimensions such as factual consistency or abstractiveness. The tool together with several pre-computed model outputs is available at https://github.com/robustness-gym/summvis.

We introduce TestCase-Eval, a new benchmark for systematic evaluation of LLMs in test-case generation. TestCase-Eval includes 500 algorithm problems and 100,000 human-crafted solutions from the Codeforces platform. It focuses on two pivotal tasks: (1) Fault Coverage, which measures how well LLM-generated test sets probe diverse input scenarios and cover a wide range of potential failure modes. (2) Fault Exposure, which evaluates whether LLMs can craft a tailored test input that reveals a specific incorrect code implementation. We provide a comprehensive assessment of 19 state-of-the-art open-source and proprietary LLMs on TestCase-Eval, offering insights into their strengths and limitations in generating effective test cases for algorithm problems.

With the fast development of COVID-19 into a global pandemic, scientists around the globe are desperately searching for effective antiviral therapeutic agents. Bridging systems biology and drug discovery, we propose a deep learning framework for conditional de novo design of antiviral candidate drugs tailored against given protein targets. First, we train a multimodal ligand--protein binding affinity model on predicting affinities of antiviral compounds to target proteins and couple this model with pharmacological toxicity predictors. Exploiting this multi-objective as a reward function of a conditional molecular generator (consisting of two VAEs), we showcase a framework that navigates the chemical space toward regions with more antiviral molecules. Specifically, we explore a challenging setting of generating ligands against unseen protein targets by performing a leave-one-out-cross-validation on 41 SARS-CoV-2-related target proteins. Using deep RL, it is demonstrated that in 35 out of 41 cases, the generation is biased towards sampling more binding ligands, with an average increase of 83% comparing to an unbiased VAE. We present a case-study on a potential Envelope-protein inhibitor and perform a synthetic accessibility assessment of the best generated molecules is performed that resembles a viable roadmap towards a rapid in-vitro evaluation of potential SARS-CoV-2 inhibitors.

While synthetic data has proven effective for improving scientific reasoning in the text domain, multimodal reasoning remains constrained by the difficulty of synthesizing scientifically rigorous images. Existing Text-to-Image (T2I) models often produce outputs that are visually plausible yet scientifically incorrect, resulting in a persistent visual-logic divergence that limits their value for downstream reasoning. Motivated by recent advances in next-generation T2I models, we conduct a systematic study of scientific image synthesis across generation paradigms, evaluation, and downstream use. We analyze both direct pixel-based generation and programmatic synthesis, and propose ImgCoder, a logic-driven framework that follows an explicit "understand - plan - code" workflow to improve structural precision. To rigorously assess scientific correctness, we introduce SciGenBench, which evaluates generated images based on information utility and logical validity. Our evaluation reveals systematic failure modes in pixel-based models and highlights a fundamental expressiveness-precision trade-off. Finally, we show that fine-tuning Large Multimodal Models (LMMs) on rigorously verified synthetic scientific images yields consistent reasoning gains, with potential scaling trends analogous to the text domain, validating high-fidelity scientific synthesis as a viable path to unlocking massive multimodal reasoning capabilities.

AI scientist systems, capable of autonomously executing the full research workflow from hypothesis generation and experimentation to paper writing, hold significant potential for accelerating scientific discovery. However, the internal workflow of these systems have not been closely examined. This lack of scrutiny poses a risk of introducing flaws that could undermine the integrity, reliability, and trustworthiness of their research outputs. In this paper, we identify four potential failure modes in contemporary AI scientist systems: inappropriate benchmark selection, data leakage, metric misuse, and post-hoc selection bias. To examine these risks, we design controlled experiments that isolate each failure mode while addressing challenges unique to evaluating AI scientist systems. Our assessment of two prominent open-source AI scientist systems reveals the presence of several failures, across a spectrum of severity, which can be easily overlooked in practice. Finally, we demonstrate that access to trace logs and code from the full automated workflow enables far more effective detection of such failures than examining the final paper alone. We thus recommend journals and conferences evaluating AI-generated research to mandate submission of these artifacts alongside the paper to ensure transparency, accountability, and reproducibility.

This report evaluates PDF-to-Markdown conversion using recent Vision-Language Models (VLMs) on challenging French documents. Document parsing is a critical step for Retrieval-Augmented Generation (RAG) pipelines, where transcription and layout errors propagate to downstream retrieval and grounding. Existing benchmarks often emphasize English or Chinese and can over-penalize benign formatting and linearization choices (e.g., line breaks, list segmentation, alternative table renderings) that are largely irrelevant for downstream use. We introduce a French-focused benchmark of difficult pages selected via model-disagreement sampling from a corpus of 60{,}000 documents, covering handwritten forms, complex layouts, dense tables, and graphics-rich pages. Evaluation is performed with unit-test-style checks that target concrete failure modes (text presence, reading order, and local table constraints) combined with category-specific normalization designed to discount presentation-only variance. Across 15 models, we observe substantially higher robustness for the strongest proprietary models on handwriting and forms, while several open-weights systems remain competitive on standard printed layouts.

Large vision-language models (VLMs) are widely getting adopted in industry and academia. In this work we build a unified framework to systematically evaluate gender-profession bias in VLMs. Our evaluation encompasses all supported inference modes of the recent VLMs, including image-to-text, text-to-text, text-to-image, and image-to-image. We construct a synthetic, high-quality dataset of text and images that blurs gender distinctions across professional actions to benchmark gender bias. In our benchmarking of recent vision-language models (VLMs), we observe that different input-output modalities result in distinct bias magnitudes and directions. We hope our work will help guide future progress in improving VLMs to learn socially unbiased representations. We will release our data and code.

Evolutions in the world, such as water pouring or ice melting, happen regardless of being observed. Video world models generate "worlds" via 2D frame observations. Can these generated "worlds" evolve regardless of observation? To probe this question, we design a benchmark to evaluate whether video world models can decouple state evolution from observation. Our benchmark, STEVO-Bench, applies observation control to evolving processes via instructions of occluder insertion, turning off the light, or specifying camera "lookaway" trajectories. By evaluating video models with and without camera control for a diverse set of naturally-occurring evolutions, we expose their limitations in decoupling state evolution from observation. STEVO-Bench proposes an evaluation protocol to automatically detect and disentangle failure modes of video world models across key aspects of natural state evolution. Analysis of STEVO-Bench results provide new insight into potential data and architecture bias of present-day video world models. Project website: https://glab-caltech.github.io/STEVOBench/. Blog: https://ziqi-ma.github.io/blog/2026/outofsight/

Meta-analyses statistically aggregate the findings of different randomized controlled trials (RCTs) to assess treatment effectiveness. Because this yields robust estimates of treatment effectiveness, results from meta-analyses are considered the strongest form of evidence. However, rigorous evidence syntheses are time-consuming and labor-intensive, requiring manual extraction of data from individual trials to be synthesized. Ideally, language technologies would permit fully automatic meta-analysis, on demand. This requires accurately extracting numerical results from individual trials, which has been beyond the capabilities of natural language processing (NLP) models to date. In this work, we evaluate whether modern large language models (LLMs) can reliably perform this task. We annotate (and release) a modest but granular evaluation dataset of clinical trial reports with numerical findings attached to interventions, comparators, and outcomes. Using this dataset, we evaluate the performance of seven LLMs applied zero-shot for the task of conditionally extracting numerical findings from trial reports. We find that massive LLMs that can accommodate lengthy inputs are tantalizingly close to realizing fully automatic meta-analysis, especially for dichotomous (binary) outcomes (e.g., mortality). However, LLMs -- including ones trained on biomedical texts -- perform poorly when the outcome measures are complex and tallying the results requires inference. This work charts a path toward fully automatic meta-analysis of RCTs via LLMs, while also highlighting the limitations of existing models for this aim.

Large language models (LLMs) operate in two fundamental learning modes - fine-tuning (FT) and in-context learning (ICL) - raising key questions about which mode yields greater language proficiency and whether they differ in their inductive biases. Prior studies comparing FT and ICL have yielded mixed and inconclusive results due to inconsistent experimental setups. To enable a rigorous comparison, we propose a formal language learning task - offering precise language boundaries, controlled string sampling, and no data contamination - and introduce a discriminative test for language proficiency, where an LLM succeeds if it assigns higher generation probability to in-language strings than to out-of-language strings. Empirically, we find that: (a) FT has greater language proficiency than ICL on in-distribution generalization, but both perform equally well on out-of-distribution generalization. (b) Their inductive biases, measured by the correlation in string generation probabilities, are similar when both modes partially learn the language but diverge at higher proficiency levels. (c) Unlike FT, ICL performance differs substantially across models of varying sizes and families and is sensitive to the token vocabulary of the language. Thus, our work demonstrates the promise of formal languages as a controlled testbed for evaluating LLMs, behaviors that are difficult to isolate in natural language datasets. Our source code is available at https://github.com/bishwamittra/formallm.

Speaker diarization is the task of partitioning audio into segments according to speaker identity, answering the question of "who spoke when" in multi-speaker conversation recordings. While diarization is an essential task for many downstream applications, it remains an unsolved problem. Errors in diarization propagate to downstream systems and cause wide-ranging failures. To this end, we examine exact failure modes by evaluating five state-of-the-art diarization models, across four diarization datasets spanning multiple languages and acoustic conditions. The evaluation datasets consist of 196.6 hours of multilingual audio, including English, Mandarin, German, Japanese, and Spanish. Overall, we find that PyannoteAI achieves the best performance at 11.2% DER, while DiariZen provides a competitive open-source alternative at 13.3% DER. When analyzing failure cases, we find that the primary cause of diarization errors stem from missed speech segments followed by speaker confusion, especially in high-speaker count settings.

The majority of data in businesses and industries is stored in tables, databases, and data warehouses. Reasoning with table-structured data poses significant challenges for large language models (LLMs) due to its hidden semantics, inherent complexity, and structured nature. One of these challenges is lacking an effective evaluation benchmark fairly reflecting the performances of LLMs on broad table reasoning abilities. In this paper, we fill in this gap, presenting a comprehensive table reasoning evolution benchmark, TReB, which measures both shallow table understanding abilities and deep table reasoning abilities, a total of 26 sub-tasks. We construct a high quality dataset through an iterative data processing procedure. We create an evaluation framework to robustly measure table reasoning capabilities with three distinct inference modes, TCoT, PoT and ICoT. Further, we benchmark over 20 state-of-the-art LLMs using this frame work and prove its effectiveness. Experimental results reveal that existing LLMs still have significant room for improvement in addressing the complex and real world Table related tasks. Both the dataset and evaluation framework are publicly available, with the dataset hosted on [HuggingFace] and the framework on [GitHub].

Retrieval-Augmented Generation (RAG) has emerged as a common paradigm to use Large Language Models (LLMs) alongside private and up-to-date knowledge bases. In this work, we address the challenges of using LLM-as-a-Judge when evaluating grounded answers generated by RAG systems. To assess the calibration and discrimination capabilities of judge models, we identify 7 generator failure modes and introduce GroUSE (Grounded QA Unitary Scoring of Evaluators), a meta-evaluation benchmark of 144 unit tests. This benchmark reveals that existing automated RAG evaluation frameworks often overlook important failure modes, even when using GPT-4 as a judge. To improve on the current design of automated RAG evaluation frameworks, we propose a novel pipeline and find that while closed models perform well on GroUSE, state-of-the-art open-source judges do not generalize to our proposed criteria, despite strong correlation with GPT-4's judgement. Our findings suggest that correlation with GPT-4 is an incomplete proxy for the practical performance of judge models and should be supplemented with evaluations on unit tests for precise failure mode detection. We further show that finetuning Llama-3 on GPT-4's reasoning traces significantly boosts its evaluation capabilities, improving upon both correlation with GPT-4's evaluations and calibration on reference situations.

Is basic visual understanding really solved in state-of-the-art VLMs? We present VisualOverload, a slightly different visual question answering (VQA) benchmark comprising 2,720 question-answer pairs, with privately held ground-truth responses. Unlike prior VQA datasets that typically focus on near global image understanding, VisualOverload challenges models to perform simple, knowledge-free vision tasks in densely populated (or, overloaded) scenes. Our dataset consists of high-resolution scans of public-domain paintings that are populated with multiple figures, actions, and unfolding subplots set against elaborately detailed backdrops. We manually annotated these images with questions across six task categories to probe for a thorough understanding of the scene. We hypothesize that current benchmarks overestimate the performance of VLMs, and encoding and reasoning over details is still a challenging task for them, especially if they are confronted with densely populated scenes. Indeed, we observe that even the best model (o3) out of 37 tested models only achieves 19.6% accuracy on our hardest test split and overall 69.5% accuracy on all questions. Beyond a thorough evaluation, we complement our benchmark with an error analysis that reveals multiple failure modes, including a lack of counting skills, failure in OCR, and striking logical inconsistencies under complex tasks. Altogether, VisualOverload exposes a critical gap in current vision models and offers a crucial resource for the community to develop better models. Benchmark: http://paulgavrikov.github.io/visualoverload

Recent growth in reinforcement learning (RL) has surfaced a need for diverse, specialized training environments. Hand-curated environments with fixed task and reward difficulties become ineffective signals as model performance improves, and sparse rewards over long horizons induce mode collapse on specific workflows or tool structures. World models that simulate environment states have matched pure rollout performance, making them promising for scaling diversity on-demand. However, autoregressive (AR) world models suffer from a left-to-right bias preventing conditioning on globally interdependent state anchors such as tool schemas, prior turns, and expected outcomes. We (i) formalize text-based world modeling as a steerable transition-dynamics problem decomposed into initial state, task context, tool schemas, domain rules, and steering directives, and (ii) curate 239,403 grounded state-action trajectories spanning nine open-source environments and twelve frontier model families. We compare AR LMs and masked diffusion language models (MDLMs), showing MDLMs, via bidirectional anchor-aware denoising, achieve better coherence, groundedness, and empirically validated rollout diversity than LLMs over 4x their parameter size, at comparable inference latency. We introduce a plug-and-play GRPO training framework with deterministic state checks, and perform zero-shot transfer ablations on three OOD environments (ScienceWorld, ALFWorld, AppWorld) across three 1.2B-7B agent backbones (LFM2.5, Qwen3, Mistral), achieving up to 47% absolute gains over baselines without environment-specific fine-tuning. We further conduct behavioral analysis of failure modes under adversarial scenarios and human evaluation on realism, outcome correctness, and training utility. We open-source our work to encourage research in this direction.

Significant progress has been made in the field of Instruction-based Image Editing Models (IIEMs). However, while these models demonstrate plausible adherence to instructions and strong reasoning ability on current benchmarks, their ability to edit small objects remains underexplored, despite its importance for precise local editing and refining details in both real and generated images. In this paper, we introduce DeepLookEditBench (DLEBench), the first benchmark dedicated to assessing the abilities of IIEMs in editing small-scale objects. Specifically, we construct a challenging testbed comprising 1889 samples across seven instruction types. In these samples, target objects occupy only 1%-10% of the image area, covering complex scenarios such as partial occlusion and multi-object editing. To ensure robust evaluation on this benchmark, we propose an evaluation protocol with refined score rubrics to minimize subjectivity and ambiguity in two criteria: Instruction Following and Visual Consistency. This protocol also introduces a dual-mode evaluation framework (Tool-driven and Oracle-guided Modes) addressing the misalignment between LMM-as-a-Judge and human judgements on DLEBench. Empirical results on 10 IIEMs reveal significant performance gaps in small-scale object editing, highlighting the need for specialized benchmarks to advance this ability.

Vision-language models (VLMs) are increasingly used to detect whether AI-generated images contain visible artifacts, yet their ability to analyze such artifacts remains poorly understood. A correct image-level decision can still hide important failures: a model may correctly flag an artifact while relying on the wrong visual cue, selecting the wrong region, or describing a defect that the image does not support. To evaluate these behaviors directly, we introduce SalArt-VQA, a diagnostic benchmark for fine-grained SALient ARTifact understanding in AI-generated images. SalArt-VQA contains 950 images and 3,681 human-authored multiple-choice questions spanning artifact images, matched real reference images, and paired generated reference images. Four aligned question types evaluate presence detection, semantic localization, spatial grounding, and evidence-grounded defect identification, while the reference splits test calibration and abstention when the annotated defect is absent. Across 20 VLMs, SalArt-VQA reveals failures that image-level detection accuracy hides: the strongest model reaches 99.37% detection recall on artifact images but answers all four artifact-side questions correctly on only 53.26% of images. Comparing artifact images with artifact-free references reveals a sensitivity-calibration tradeoff: sensitive models often make unsupported artifact claims, while conservative models avoid false alarms largely by missing real artifacts. These results show that high artifact detection accuracy alone does not imply grounded artifact understanding. SalArt-VQA exposes these hidden failure modes and provides a fine-grained evaluation of whether VLM artifact claims are supported by local visual evidence.

We introduce AInsteinBench, a large-scale benchmark for evaluating whether large language model (LLM) agents can operate as scientific computing development agents within real research software ecosystems. Unlike existing scientific reasoning benchmarks which focus on conceptual knowledge, or software engineering benchmarks that emphasize generic feature implementation and issue resolving, AInsteinBench evaluates models in end-to-end scientific development settings grounded in production-grade scientific repositories. The benchmark consists of tasks derived from maintainer-authored pull requests across six widely used scientific codebases, spanning quantum chemistry, quantum computing, molecular dynamics, numerical relativity, fluid dynamics, and cheminformatics. All benchmark tasks are carefully curated through multi-stage filtering and expert review to ensure scientific challenge, adequate test coverage, and well-calibrated difficulty. By leveraging evaluation in executable environments, scientifically meaningful failure modes, and test-driven verification, AInsteinBench measures a model's ability to move beyond surface-level code generation toward the core competencies required for computational scientific research.

We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate nine versions of GPT models, including ChatGPT and GPT-4. We show that our method for translation quality assessment only works with GPT~3.5 and larger models. Comparing to results from WMT22's Metrics shared task, our method achieves state-of-the-art accuracy in both modes when compared to MQM-based human labels. Our results are valid on the system level for all three WMT22 Metrics shared task language pairs, namely English into German, English into Russian, and Chinese into English. This provides a first glimpse into the usefulness of pre-trained, generative large language models for quality assessment of translations. We publicly release all our code and prompt templates used for the experiments described in this work, as well as all corresponding scoring results, to allow for external validation and reproducibility.

Many recent advancements in Computer Vision are attributed to large datasets. Open-source software packages for Machine Learning and inexpensive commodity hardware have reduced the barrier of entry for exploring novel approaches at scale. It is possible to train models over millions of examples within a few days. Although large-scale datasets exist for image understanding, such as ImageNet, there are no comparable size video classification datasets. In this paper, we introduce YouTube-8M, the largest multi-label video classification dataset, composed of ~8 million videos (500K hours of video), annotated with a vocabulary of 4800 visual entities. To get the videos and their labels, we used a YouTube video annotation system, which labels videos with their main topics. While the labels are machine-generated, they have high-precision and are derived from a variety of human-based signals including metadata and query click signals. We filtered the video labels (Knowledge Graph entities) using both automated and manual curation strategies, including asking human raters if the labels are visually recognizable. Then, we decoded each video at one-frame-per-second, and used a Deep CNN pre-trained on ImageNet to extract the hidden representation immediately prior to the classification layer. Finally, we compressed the frame features and make both the features and video-level labels available for download. We trained various (modest) classification models on the dataset, evaluated them using popular evaluation metrics, and report them as baselines. Despite the size of the dataset, some of our models train to convergence in less than a day on a single machine using TensorFlow. We plan to release code for training a TensorFlow model and for computing metrics.

Voice agents, artificial intelligence systems that conduct spoken conversations to complete tasks, are increasingly deployed across enterprise applications. However, no existing benchmark jointly addresses two core evaluation challenges: generating realistic simulated conversations, and measuring quality across the full scope of voice-specific failure modes. We present EVA-Bench, an end-to-end evaluation framework that addresses both. On the simulation side, EVA-Bench orchestrates bot-to-bot audio conversations over dynamic multi-turn dialogues, with automatic simulation validation that detects user simulator error and appropriately regenerates conversations before scoring. On the measurement side, EVA-Bench introduces two composite metrics: EVA-A (Accuracy), capturing task completion, faithfulness, and audio-level speech fidelity; and EVA-X (Experience), capturing conversation progression, spoken conciseness, and turn-taking timing. Both metrics apply to different agent architectures, enabling direct cross-architecture comparison. EVA-Bench includes 213 scenarios across three enterprise domains, a controlled perturbation suite for accent and noise robustness, and pass@1, pass@k, pass^k measurements that distinguish peak from reliable capability. Across 12 systems spanning all three architectures, we find: (1) no system simultaneously exceeds 0.5 on both EVA-A pass@1 and EVA-X pass@1; (2) peak and reliable performance diverge substantially (median pass@k - pass^k gap of 0.44 on EVA-A); and (3) accent and noise perturbations expose substantial robustness gaps, with effects varying across architectures, systems, and metrics (mean up to 0.314). We release the full framework, evaluation suite, and benchmark data under an open-source license.

Community Notes, the crowd-sourced misinformation governance system on X (formerly Twitter), allows users to flag misleading posts, attach contextual notes, and rate the notes' helpfulness. However, our empirical analysis of 30.8K health-related notes reveals substantial latency, with a median delay of 17.6 hours before notes receive a helpfulness status. To improve responsiveness during real-world misinformation surges, we propose CrowdNotes+, a unified LLM-based framework that augments Community Notes for faster and more reliable health misinformation governance. CrowdNotes+ integrates two modes: (1) evidence-grounded note augmentation and (2) utility-guided note automation, supported by a hierarchical three-stage evaluation of relevance, correctness, and helpfulness. We instantiate the framework with HealthNotes, a benchmark of 1.2K health notes annotated for helpfulness, and a fine-tuned helpfulness judge. Our analysis first uncovers a key loophole in current crowd-sourced governance: voters frequently conflate stylistic fluency with factual accuracy. Addressing this via our hierarchical evaluation, experiments across 15 representative LLMs demonstrate that CrowdNotes+ significantly outperforms human contributors in note correctness, helpfulness, and evidence utility.

Despite widespread deployment of Large Language Models, systematic evaluation of instruction-following capabilities remains challenging. While comprehensive benchmarks exist, focused assessments that quickly diagnose specific instruction adherence patterns are valuable. As newer models may be trained on existing benchmarks, novel evaluation approaches are needed to assess genuine capabilities rather than memorized performance. This paper presents a streamlined evaluation framework using twenty carefully designed prompts to assess LLM instruction-following across diverse task categories. We demonstrate this framework through a large-scale empirical study conducted on October 14, 2025, testing 256 verified working models from 331 available via OpenRouter. To ensure methodological rigor and prevent selection bias, we first verified each model's basic functionality before inclusion. Unlike large-scale benchmarks requiring extensive computational resources, our approach offers a practical diagnostic tool researchers and practitioners can readily apply. Our methodology builds upon verifiable instructions while introducing a compact test suite balancing comprehensiveness with efficiency. Each prompt targets distinct aspects of instruction following, including format compliance, content constraints, logical sequencing, and multi-step task execution. We evaluate models from major providers (OpenAI, Anthropic, Google, Meta, Mistral) and emerging implementations (Qwen, DeepSeek, community models), providing comparative performance analysis. Our findings reveal consistent failure modes and identify specific instruction types posing particular challenges. This work contributes both a practical evaluation tool and one of the most comprehensive empirical analyses of instruction-following capabilities across the contemporary LLM landscape.

End-to-end autonomous driving models are typically trained on multi-city datasets using supervised ImageNet-pretrained backbones, yet their ability to generalize to unseen cities remains largely unexamined. When training and evaluation data are geographically mixed, models may implicitly rely on city-specific cues, masking failure modes that would occur under real-world domain shifts when generalizing to new locations. In this work, we formulate zero-shot cross-city transfer as a controlled representation-level stress test for end-to-end autonomous driving and ask how visual pretraining affects transfer behavior under geographic domain shift. We conduct a comprehensive study by integrating self-supervised backbones I-JEPA, DINOv2, and MAE into planning frameworks. We evaluate performance under strict geographic splits on nuScenes in the open-loop setting and on NAVSIM in the closed-loop evaluation protocol. Our experiments reveal a substantial generalization gap when transferring models across cities with different road topologies, traffic conventions, and visual environments. In open-loop evaluation, a supervised backbone exhibits severe degradation when transferring between cities, yet some domain-specific self-supervised methods can substantially reduce both displacement and collision degradation. In closed-loop evaluation, self-supervised pretraining improves average out-of-distribution PDMS in several single-city training settings. Our results provide empirical evidence that representation learning influences the robustness of cross-city planning and motivate zero-shot geographic transfer as an important stress test for evaluating end-to-end autonomous driving systems.

The behavior and decision making of groups or communities can be dramatically influenced by individuals pushing particular agendas, e.g., to promote or disparage a person or an activity, to call for action, etc.. In the examination of online influence campaigns, particularly those related to important political and social events, scholars often concentrate on identifying the sources responsible for setting and controlling the agenda (e.g., public media). In this article we present a methodology for detecting specific instances of agenda control through social media where annotated data is limited or non-existent. By using a modest corpus of Twitter messages centered on the 2022 French Presidential Elections, we carry out a comprehensive evaluation of various approaches and techniques that can be applied to this problem. Our findings demonstrate that by treating the task as a textual entailment problem, it is possible to overcome the requirement for a large annotated training dataset.

We present NUPunkt and CharBoundary, two sentence boundary detection libraries optimized for high-precision, high-throughput processing of legal text in large-scale applications such as due diligence, e-discovery, and legal research. These libraries address the critical challenges posed by legal documents containing specialized citations, abbreviations, and complex sentence structures that confound general-purpose sentence boundary detectors. Our experimental evaluation on five diverse legal datasets comprising over 25,000 documents and 197,000 annotated sentence boundaries demonstrates that NUPunkt achieves 91.1% precision while processing 10 million characters per second with modest memory requirements (432 MB). CharBoundary models offer balanced and adjustable precision-recall tradeoffs, with the large model achieving the highest F1 score (0.782) among all tested methods. Notably, NUPunkt provides a 29-32% precision improvement over general-purpose tools while maintaining exceptional throughput, processing multi-million document collections in minutes rather than hours. Both libraries run efficiently on standard CPU hardware without requiring specialized accelerators. NUPunkt is implemented in pure Python with zero external dependencies, while CharBoundary relies only on scikit-learn and optional ONNX runtime integration for optimized performance. Both libraries are available under the MIT license, can be installed via PyPI, and can be interactively tested at https://sentences.aleainstitute.ai/. These libraries address critical precision issues in retrieval-augmented generation systems by preserving coherent legal concepts across sentences, where each percentage improvement in precision yields exponentially greater reductions in context fragmentation, creating cascading benefits throughout retrieval pipelines and significantly enhancing downstream reasoning quality.

As AI systems move from generating text to accomplishing goals through sustained interaction, the ability to model environment dynamics becomes a central bottleneck. Agents that manipulate objects, navigate software, coordinate with others, or design experiments require predictive environment models, yet the term world model carries different meanings across research communities. We introduce a "levels x laws" taxonomy organized along two axes. The first defines three capability levels: L1 Predictor, which learns one-step local transition operators; L2 Simulator, which composes them into multi-step, action-conditioned rollouts that respect domain laws; and L3 Evolver, which autonomously revises its own model when predictions fail against new evidence. The second identifies four governing-law regimes: physical, digital, social, and scientific. These regimes determine what constraints a world model must satisfy and where it is most likely to fail. Using this framework, we synthesize over 400 works and summarize more than 100 representative systems spanning model-based reinforcement learning, video generation, web and GUI agents, multi-agent social simulation, and AI-driven scientific discovery. We analyze methods, failure modes, and evaluation practices across level-regime pairs, propose decision-centric evaluation principles and a minimal reproducible evaluation package, and outline architectural guidance, open problems, and governance challenges. The resulting roadmap connects previously isolated communities and charts a path from passive next-step prediction toward world models that can simulate, and ultimately reshape, the environments in which agents operate.

Modern analyst agents must reason over complex, high token inputs, including dozens of retrieved documents, tool outputs, and time sensitive data. While prior work has produced tool calling benchmarks and examined factuality in knowledge augmented systems, relatively little work studies their intersection: settings where LLMs must integrate large volumes of dynamic, structured and unstructured multi tool outputs. We investigate LLM failure modes in this regime using crypto as a representative high data density domain. We introduce (1) CryptoAnalystBench, an analyst aligned benchmark of 198 production crypto and DeFi queries spanning 11 categories; (2) an agentic harness equipped with relevant crypto and DeFi tools to generate responses across multiple frontier LLMs; and (3) an evaluation pipeline with citation verification and an LLM as a judge rubric spanning four user defined success dimensions: relevance, temporal relevance, depth, and data consistency. Using human annotation, we develop a taxonomy of seven higher order error types that are not reliably captured by factuality checks or LLM based quality scoring. We find that these failures persist even in state of the art systems and can compromise high stakes decisions. Based on this taxonomy, we refine the judge rubric to better capture these errors. While the judge does not align with human annotators on precise scoring across rubric iterations, it reliably identifies critical failure modes, enabling scalable feedback for developers and researchers studying analyst style agents. We release CryptoAnalystBench with annotated queries, the evaluation pipeline, judge rubrics, and the error taxonomy, and outline mitigation strategies and open challenges in evaluating long form, multi tool augmented systems.

Recent advances in large language models have enabled natural language to serve as a flexible interface for controlling complex systems, but often at the cost of large-scale multimodal training or weakened domain-specific inductive biases. In structured decision-making domains such as chess, specialized policy networks achieve strong performance but lack semantic controllability, while prompt-conditioned language models are more flexible yet typically exhibit weaker domain grounding. We propose UniMaia, a framework for prompt-conditioned policy modulation that adapts a frozen Lc0-based chess policy network using a parameter-efficient text encoder and a ControlNet-style conditioning mechanism. UniMaia enables semantic control over gameplay, including opening selection and player strength, while preserving the pretrained policy representations. We further introduce UniMaia-Aux, which incorporates auxiliary temporal conditioning and behavioral prediction objectives. To support this work, we construct a large-scale metadata-augmented Lichess dataset, develop a semi-automated prompt-generation pipeline, and introduce benchmarks spanning both prompt-conditioned and metadata-conditioned settings. UniMaia achieves state-of-the-art expected accuracy on several prompt-conditioned benchmarks and competitive top-move accuracy on general instruction-following tasks, while remaining competitive with dedicated metadata-conditioned approaches on human move prediction benchmarks. UniMaia-Aux further improves expected accuracy and behavioral modeling across several evaluation settings, with modest trade-offs in top-move accuracy. Overall, our results demonstrate that prompt-conditioned control of domain-specific policy networks is feasible without end-to-end multimodal training, while highlighting trade-offs between controllability and predictive performance.

Chemical reasoning inherently integrates visual, textual, and symbolic modalities, yet existing benchmarks rarely capture this complexity, often relying on simple image-text pairs with limited chemical semantics. As a result, the actual ability of Multimodal Large Language Models (MLLMs) to process and integrate chemically meaningful information across modalities remains unclear. We introduce ChemVTS-Bench, a domain-authentic benchmark designed to systematically evaluate the Visual-Textual-Symbolic (VTS) reasoning abilities of MLLMs. ChemVTS-Bench contains diverse and challenging chemical problems spanning organic molecules, inorganic materials, and 3D crystal structures, with each task presented in three complementary input modes: (1) visual-only, (2) visual-text hybrid, and (3) SMILES-based symbolic input. This design enables fine-grained analysis of modality-dependent reasoning behaviors and cross-modal integration. To ensure rigorous and reproducible evaluation, we further develop an automated agent-based workflow that standardizes inference, verifies answers, and diagnoses failure modes. Extensive experiments on state-of-the-art MLLMs reveal that visual-only inputs remain challenging, structural chemistry is the hardest domain, and multimodal fusion mitigates but does not eliminate visual, knowledge-based, or logical errors, highlighting ChemVTS-Bench as a rigorous, domain-faithful testbed for advancing multimodal chemical reasoning. All data and code will be released to support future research.

We introduce self-invoking code generation, a new task designed to evaluate the progressive reasoning and problem-solving capabilities of LLMs. In this task, models are presented with a base problem and a related, more complex problem. They must solve the base problem and then utilize its solution to address the more complex one. This work features three key contributions. First, we propose a general recipe for generating more challenging versions of existing benchmarks, resulting in three new benchmarks: HumanEval Pro, MBPP Pro, and BigCodeBench-Lite Pro, specifically designed to assess LLMs on self-invoking code generation. Second, from the analysis of experimental results over twenty LLMs on our benchmarks, we have two important observations: (i) Most LLMs excel in traditional code generation benchmarks like HumanEval and MBPP, but their performance declines on self-invoking tasks. For example, o1-mini achieves 96.2% pass@1 on HumanEval but only 76.2% on HumanEval Pro. (ii) On self-invoking code generation task, the instruction-tuned models demonstrate only marginal improvements compared to the base models. Third, we disclose the types of failure modes that exist in our evaluation results. All these results underscore the need for further advancements in self-invoking code generation tasks and provide a new direction for future research on enhancing LLMs' code reasoning capabilities.

Entity relationship classification remains a challenging task in information extraction, especially in scenarios with limited labeled data and complex relational structures. In this study, we conduct a comparative analysis of three distinct AI agent architectures designed to perform relation classification using large language models (LLMs). The agentic architectures explored include (1) reflective self-evaluation, (2) hierarchical task decomposition, and (3) a novel multi-agent dynamic example generation mechanism, each leveraging different modes of reasoning and prompt adaptation. In particular, our dynamic example generation approach introduces real-time cooperative and adversarial prompting. We systematically compare their performance across multiple domains and model backends. Our experiments demonstrate that multi-agent coordination consistently outperforms standard few-shot prompting and approaches the performance of fine-tuned models. These findings offer practical guidance for the design of modular, generalizable LLM-based systems for structured relation extraction. The source codes and dataset are available at https://github.com/maryambrj/ALIEN.git.

Observational studies can yield clinically actionable evidence at scale, but executing them on real-world databases is open-ended and requires coherent decisions across cohort construction, analysis, and reporting. Prior evaluations of LLM agents emphasize isolated steps or single answers, missing the integrity and internal structure of the resulting evidence bundle. To address this gap, we introduce RWE-bench, a benchmark grounded in MIMIC-IV and derived from peer-reviewed observational studies. Each task provides the corresponding study protocol as the reference standard, requiring agents to execute experiments in a real database and iteratively generate tree-structured evidence bundles. We evaluate six LLMs (three open-source, three closed-source) under three agent scaffolds using both question-level correctness and end-to-end task metrics. Across 162 tasks, task success is low: the best agent reaches 39.9%, and the best open-source model reaches 30.4%. Agent scaffolds also matter substantially, causing over 30% variation in performance metrics. Furthermore, we implement an automated cohort evaluation method to rapidly localize errors and identify agent failure modes. Overall, the results highlight persistent limitations in agents' ability to produce end-to-end evidence bundles, and efficient validation remains an important direction for future work. Code and data are available at https://github.com/somewordstoolate/RWE-bench.

Detecting prompt injection and jailbreak attacks is critical for deploying LLM-based agents safely. As agents increasingly process untrusted data from emails, documents, tool outputs, and external APIs, robust attack detection becomes essential. Yet current evaluation practices and production systems have fundamental limitations. We present a comprehensive analysis using a diverse benchmark of 18 datasets spanning harmful requests, jailbreaks, indirect prompt injections, and extraction attacks. We propose Leave-One-Dataset-Out (LODO) evaluation to measure true out-of-distribution generalization, revealing that the standard practice of train-test splits from the same dataset sources severely overestimates performance: aggregate metrics show an 8.4 percentage point AUC inflation, but per-dataset gaps range from 1% to 25% accuracy-exposing heterogeneous failure modes. To understand why classifiers fail to generalize, we analyze Sparse Auto-Encoder (SAE) feature coefficients across LODO folds, finding that 28% of top features are dataset-dependent shortcuts whose class signal depends on specific dataset compositions rather than semantic content. We systematically compare production guardrails (PromptGuard 2, LlamaGuard) and LLM-as-judge approaches on our benchmark, finding all three fail on indirect attacks targeting agents (7-37% detection) and that PromptGuard 2 and LlamaGuard cannot evaluate agentic tool injection due to architectural limitations. Finally, we show that LODO-stable SAE features provide more reliable explanations for classifier decisions by filtering dataset artifacts. We release our evaluation framework at https://github.com/maxf-zn/prompt-mining to establish LODO as the appropriate protocol for prompt attack detection research.

This study presents a comparative evaluation of 22 large language models LLMs on the Spanish Medical Intern Resident MIR examinations for 2024 and 2025 with a focus on clinical reasoning domain specific expertise and multimodal processing capabilities The MIR exam consisting of 210 multiple choice questions some requiring image interpretation serves as a stringent benchmark for assessing both factual recall and complex clinical problem solving skills Our investigation encompasses general purpose models such as GPT4 Claude LLaMA and Gemini as well as specialized fine tuned systems like Miri Pro which leverages proprietary Spanish healthcare data to excel in medical contexts Recent market entries Deepseek and Grok have further enriched the evaluation landscape particularly for tasks that demand advanced visual and semantic analysis The findings indicate that while general purpose LLMs perform robustly overall fine tuned models consistently achieve superior accuracy especially in addressing nuanced domain specific challenges A modest performance decline observed between the two exam cycles appears attributable to the implementation of modified questions designed to mitigate reliance on memorization The results underscore the transformative potential of domain specific fine tuning and multimodal integration in advancing medical AI applications They also highlight critical implications for the future integration of LLMs into medical education training and clinical decision making emphasizing the importance of balancing automated reasoning with ethical and context aware judgment

This work presents BAdam, an optimizer that leverages the block coordinate optimization framework with Adam as the inner solver. BAdam offers a memory efficient approach to the full parameter finetuning of large language models and reduces running time of the backward process thanks to the chain rule property. Experimentally, we apply BAdam to instruction-tune the Llama 2-7B model on the Alpaca-GPT4 dataset using a single RTX3090-24GB GPU. The results indicate that BAdam exhibits superior convergence behavior in comparison to LoRA and LOMO. Furthermore, our downstream performance evaluation of the instruction-tuned models using the MT-bench shows that BAdam modestly surpasses LoRA and more substantially outperforms LOMO. Finally, we compare BAdam with Adam on a medium-sized task, i.e., finetuning RoBERTa-large on the SuperGLUE benchmark. The results demonstrate that BAdam is capable of narrowing the performance gap with Adam. Our code is available at https://github.com/Ledzy/BAdam.

We introduce WorldOlympiad, a benchmark for diagnosing video-based world models across physical faithfulness, geometric consistency, and interaction fidelity. While existing benchmarks often focus on visual quality, semantic alignment, or short-term temporal coherence, they provide limited insight into whether generated videos obey physical rules, preserve coherent 3D structure, and sustain controllable interactions over long horizons. To address this gap, WorldOlympiad decomposes world-model evaluation into three complementary dimensions. The physical track uses object segmentation and MLLM-as-judge to assess whether generated videos follow interpretable rules in mechanics, thermal phenomena, and material properties. The geometry track reconstructs generated videos with Gaussian splatting and evaluates structural consistency, cross-view coherence, and camera-trajectory alignment. The interaction track assesses whether generated rollouts follow complex action prompts and maintain smooth, coherent transitions across consecutive video chunks. WorldOlympiad further covers three major downstream scenarios, including gaming, robotics, and general real-world videos, capturing diverse challenges from interactive control and embodied manipulation to open-domain motion and camera dynamics. Together, these tracks and scenarios form a scalable and interpretable evaluation suite that exposes failure modes beyond generic video quality. Experiments on state-of-the-art models reveal substantial gaps in physical reasoning, 3D consistency, and long-horizon interaction, underscoring the need for more structured evaluation protocols for generative world models.

GUI agents, powered by large foundation models, can interact with digital interfaces, enabling various applications in web automation, mobile navigation, and software testing. However, their increasing autonomy has raised critical concerns about their security, privacy, and safety. This survey examines the trustworthiness of GUI agents in five critical dimensions: security vulnerabilities, reliability in dynamic environments, transparency and explainability, ethical considerations, and evaluation methodologies. We also identify major challenges such as vulnerability to adversarial attacks, cascading failure modes in sequential decision-making, and a lack of realistic evaluation benchmarks. These issues not only hinder real-world deployment but also call for comprehensive mitigation strategies beyond task success. As GUI agents become more widespread, establishing robust safety standards and responsible development practices is essential. This survey provides a foundation for advancing trustworthy GUI agents through systematic understanding and future research.

Large language models trained for safety and harmlessness remain susceptible to adversarial misuse, as evidenced by the prevalence of "jailbreak" attacks on early releases of ChatGPT that elicit undesired behavior. Going beyond recognition of the issue, we investigate why such attacks succeed and how they can be created. We hypothesize two failure modes of safety training: competing objectives and mismatched generalization. Competing objectives arise when a model's capabilities and safety goals conflict, while mismatched generalization occurs when safety training fails to generalize to a domain for which capabilities exist. We use these failure modes to guide jailbreak design and then evaluate state-of-the-art models, including OpenAI's GPT-4 and Anthropic's Claude v1.3, against both existing and newly designed attacks. We find that vulnerabilities persist despite the extensive red-teaming and safety-training efforts behind these models. Notably, new attacks utilizing our failure modes succeed on every prompt in a collection of unsafe requests from the models' red-teaming evaluation sets and outperform existing ad hoc jailbreaks. Our analysis emphasizes the need for safety-capability parity -- that safety mechanisms should be as sophisticated as the underlying model -- and argues against the idea that scaling alone can resolve these safety failure modes.

Large language model (LLM) agents are increasingly used to assist with operations research (OR) modeling, yet existing OR-oriented benchmarks often reduce evaluation to one-shot translation from a self-contained problem statement into a mathematical formulation or solver program. Such settings abstract away two characteristics of real industrial OR workflows: persistent multi-artifact workspaces and multi-stage task lifecycles. We introduce OR-Space, a full-lifecycle workspace benchmark for evaluating industrial optimization agents across model construction, model revision, and grounded explanation. Each instance is an executable workspace containing business documents, structured data, optional code artifacts, solver outputs, and task-specific evaluators distributed across interdependent files. OR-Space defines three task modes: Build, where agents construct solver-ready optimization models from heterogeneous artifacts; Revise, where agents modify existing models under changing requirements or solver feedback while preserving valid prior logic; and Explain, where agents answer grounded questions about solutions, constraints, and business implications using evidence spread across workspace artifacts. By combining persistent workspaces with lifecycle-oriented tasks, OR-Space evaluates whether agents can perform reliable optimization work beyond end-to-end text generation. We describe the benchmark design, evaluation protocol, and quality-control pipeline, and position OR-Space as a benchmark for studying the reliability, failure modes, and practical readiness of LLM agents in industrial OR workflows.

We present a radiology-specific multimodal model for the task for generating radiological reports from chest X-rays (CXRs). Our work builds on the idea that large language model(s) can be equipped with multimodal capabilities through alignment with pre-trained vision encoders. On natural images, this has been shown to allow multimodal models to gain image understanding and description capabilities. Our proposed model (MAIRA-1) leverages a CXR-specific image encoder in conjunction with a fine-tuned large language model based on Vicuna-7B, and text-based data augmentation, to produce reports with state-of-the-art quality. In particular, MAIRA-1 significantly improves on the radiologist-aligned RadCliQ metric and across all lexical metrics considered. Manual review of model outputs demonstrates promising fluency and accuracy of generated reports while uncovering failure modes not captured by existing evaluation practices. More information and resources can be found on the project website: https://aka.ms/maira.

Given a database S and a natural language question Q, text-to-SQL systems aim to generate an SQL query that correctly answers Q when executed against S. Currently, popular text-to-SQL benchmarks mostly assume unrestricted access to S; in practice, however, user access is often restricted, e.g., through role-based access control (RBAC) policies. This leads to a potential disconnect between benchmarking results and real-world performance: an LLM with high benchmark scores might perform poorly in an access-controlled environment, by frequently violating RBAC, or rejecting a query q that could be answered with only permitted data in S. Motivated by this, we present a comprehensive text-to-SQL benchmarking framework with realistic RBAC constraints, which features an LLM-assisted workflow that augments existing text-to-SQL benchmarks with plausible user roles and access policies. To do so, we formulate the problem of role synthesis as a structured reasoning process over the database schema, in which the LLM first infers the application context from the schema, and then derives role responsibilities and access scopes consistent with this context. This process is audited by human-in-the-loop quality control, in which domain experts perform metric-guided screening on the generated roles. Besides the augmented dataset, the proposed framework also contains evaluation metrics that identify RBAC-specific failure modes, and disentangle SQL utility from access-control compliance. We apply the proposed framework to several widely-used benchmarks, and conduct a systematic empirical study of state-of-the-art text-to-SQL systems. The results show that many solutions (especially open-weight LLMs) with high benchmarking scores under an unrestricted setting suffer sharp performance degradation once access constraints are in place, due to frequent RBAC violations.

With LLMs shifting their role from statistical modeling of language to serving as general-purpose AI agents, how should LLM evaluations change? Arguably, a key ability of an AI agent is to flexibly combine, as needed, the basic skills it has learned. The capability to combine skills plays an important role in (human) pedagogy and also in a paper on emergence phenomena (Arora & Goyal, 2023). This work introduces Skill-Mix, a new evaluation to measure ability to combine skills. Using a list of N skills the evaluator repeatedly picks random subsets of k skills and asks the LLM to produce text combining that subset of skills. Since the number of subsets grows like N^k, for even modest k this evaluation will, with high probability, require the LLM to produce text significantly different from any text in the training set. The paper develops a methodology for (a) designing and administering such an evaluation, and (b) automatic grading (plus spot-checking by humans) of the results using GPT-4 as well as the open LLaMA-2 70B model. Administering a version of to popular chatbots gave results that, while generally in line with prior expectations, contained surprises. Sizeable differences exist among model capabilities that are not captured by their ranking on popular LLM leaderboards ("cramming for the leaderboard"). Furthermore, simple probability calculations indicate that GPT-4's reasonable performance on k=5 is suggestive of going beyond "stochastic parrot" behavior (Bender et al., 2021), i.e., it combines skills in ways that it had not seen during training. We sketch how the methodology can lead to a Skill-Mix based eco-system of open evaluations for AI capabilities of future models.