Search research

Can we trust evaluation scores to capture an LLM's true real-world performance? Certifiable evaluation answers this question by providing guarantee for LLM evaluation. In particular, existing methods sequentially curate evaluation samples and keep updating confidence intervals (CIs) that cover the true performance with high probability (e.g., 95%) until some conditions are satisfied, e.g., the CI width reaches a target precision. However, existing methods are not generally anytime-valid: the claimed coverage (e.g., 95%) may fail when CIs are repeatedly updated and used to decide when to stop, leaving a gap between theoretical rigor and practice. This paper bridges this gap by proposing Celeus, a Certifiable framework for Efficient LLM evaluation, which leverages E-processes to build anytime-valid CIs. Concretely, we propose signals that combine two ingredients: (i) Uncertainty-guided sampling to select informative samples for evaluation, and (ii) Surrogate-assisted approximations for unevaluated samples. We prove that such signals remain unbiased for the evaluation score conditional on the past, enabling statistically-grounded and anytime-valid $e$-process CIs. More importantly, the two ingredients reduce estimation variance and help reach the target precision with fewer evaluated samples. We also prove that CIs obtained by Celeus can shrink at a near-parametric rate up to logarithmic factors and analyze the oracle variance-optimal sampling rule that motivates the empirical uncertainty-guided one. Experiments show that Celeus reaches the target precision using 54-62% fewer evaluated samples than baselines, while preserving anytime-valid coverage.

Recent advances in 3D generation have led to substantial improvements in realism, controllability, and efficiency, yet the evaluation of 3D assets remains underexplored. Existing evaluation paradigms, including human evaluation, learned metrics, and vision-language models (VLMs) as judges, suffer from limitations in cost, scalability, resolution handling, or task-specific alignment. In this work, we focus on 3D mesh evaluation and introduce DB-3DME, the Dataset and Benchmark for 3D Mesh Evaluation. DB-3DME contains 2,619 synthetic 3D meshes paired with human ratings on Geometry and Prompt Adherence. Using this dataset, we systematically benchmark state-of-the-art VLMs and identify visual encoding of 3D representations as a key factor for human-aligned evaluation performance. Motivated by this finding, we fine-tune an open-weight VLM, Qwen-2.5-VL-7B, for 3D mesh evaluation by adapting the visual encoder while freezing the language model. The fine-tuned model substantially outperforms existing pre-trained VLMs across multiple evaluation dimensions, establishing a new benchmark for automatic 3D mesh evaluation. We publicly release the benchmark dataset on GitHub and Hugging Face to facilitate future research.

Knowledge graph completion (KGC) aims to predict missing facts from an observed knowledge graph (KG), playing a crucial role in a wide range of real-world applications such as drug discovery, recommender systems, and retrieval-augmented generation (RAG). Although numerous KGC models have been proposed, the evaluation of KGC remains underexplored, despite its critical role in reliably assessing model performance and selecting appropriate models for real-world applications. In this paper, we introduce two important perspectives for KGC evaluation that are overlooked by existing evaluation metrics, (P1) predictive sharpness and (P2) popularity-bias robustness. To address both perspectives, we propose a generalized evaluation framework, PROBE, which consists of a rank transformer (RT) that estimates the score of each prediction based on a desired level of predictive sharpness and a rank aggregator (RA) that determines the final evaluation score by aggregating all prediction scores according to a desired level of popularity-bias robustness. We theoretically analyze PROBE by defining six key properties for reliable KGC evaluation and prove that PROBE satisfies all the properties, while existing metrics fail to satisfy some. In particular, due to the open-world nature of KGs, an evaluation metric should preserve the relative performance of KGC models even when only incomplete facts are observed. We show that PROBE better maintains such consistency, providing a more reliable estimate of intrinsic model performance than existing metrics. Extensive experiments with six KGC models on six real-world KGs reveal that existing metrics may over- or under-estimate model performance depending on different evaluation perspectives, whereas PROBE enables a more comprehensive, flexible, and consistent evaluation of KGC models.

Large Language Models (LLMs) have become increasingly adopted in daily applications, with deep research standing out as a particularly important capability. Unlike traditional question-answering (QA) tasks, deep research report generation lacks definitive ground-truth, making reward design inherently unverifiable and limiting effective reinforcement learning. Existing approaches mitigate this challenge with LLM-as-a-judge and query-dependent evaluation rubrics, but they still rely on static evaluators that cannot adapt their standards as the solver improves, leading to insufficient and eventually saturated optimization pressure. We address this limitation with a \textbf{s}elf-evolving \textbf{co}-evolutionary training framework for deep \textbf{re}search evaluation and generation (SCORE), which tightly couples an evaluator and a solver in a shared-parameter learning process. Rather than treating generation and evaluation as isolated modules, we leverage their intrinsic connection to enable joint improvement within a single shared-parameter model. To restrict this process, we introduce a meta-harness, which dynamically controls the evaluation environment based on solver performance, encouraging valid evaluation dimensions and sufficiently deep evaluator search. Extensive experiments on deep research benchmarks demonstrate consistent improvement in report generation quality, showing that co-evolving evaluation and generation is a promising direction for training open-ended research agents.

Rapid advances in audio-video (AV) generation have enabled high-fidelity synthesis with synchronized sound, particularly for human-related scenarios involving speech and interactions. Yet evaluation for AV generation remains at an early stage, with only a few coarse-grained benchmarks for human-related scenarios and relying on limited preset evaluations with generic multimodal LLMs, leading to inaccurate assessments of model capabilities. To address these issues, we introduce AVBench, a fully automated benchmark tailored for human-centric AV generation. AVBench is built on two key designs for comprehensive and accurate evaluation: (i) Human-centric and fine-grained metrics. AVBench integrates ten evaluation dimensions designed for human-centered real-world scenarios, covering visual quality, audio quality, and multi-level consistency across modalities. These practical metrics capture human-related details that existing benchmarks often overlook. (ii) Specialized evaluators via preference learning. To address the lack of specialized training data, we construct large-scale supervision by transforming real-world videos into diverse training pairs with controlled perturbations. After fine-tuning on this high-quality dataset, the evaluators learn to reliably detect subtle cross-modal inconsistencies. Crucially, instead of producing discrete textual judgment, AVBench derives continuous evaluation scores from the model's prediction confidence on binary decisions. This probabilistic scoring mechanism enables a more reliable assessment than traditional VQA-style evaluation and aligns closely with human judgment. Taken together, AVBench offers automated evaluation for AV generation, demonstrates strong potential for data filtering, and serves as a differentiable reward signal for Reinforcement Learning from Human Feedback (RLHF).

AI evaluation is undergoing a structural change. Large language models (LLMs) are increasingly deployed as systems that act over time through tools, environments, users, and other agents, while many evaluation practices still inherit assumptions from response-centered benchmarks (e.g., fixed inputs, isolated outputs, and outcome judgments that can be made from a single response). The field has begun to build interactive benchmarks, but the resulting landscape is fragmented: benchmarks differ in what interaction artifacts they admit, how trajectories are scored, and what claims their results support. This position paper argues that interactive evaluation should be treated as a principled evaluation paradigm, not merely a new family of agent benchmarks. Simply adopting previous evaluation paradigms does not suffice. We define evaluation as an autonomous mapping from evidence to judgments, and show that interactive evaluation changes both sides of this mapping: the evidence becomes interaction-generated trajectories, while the evaluation procedure must assess process, recoverability, coordination, robustness, and system-level performance. Building on this definition, we propose a two-axis taxonomy, derive design principles and reporting standards, examine representative scenarios, and analyze how longstanding evaluation challenges reappear at the trajectory level.

Every AI benchmark operationalizes theoretical assumptions about the capability it claims to assess. When assumptions function as unexamined commitments, benchmarks stabilize the dominant paradigm by narrowing what counts as progress. Over time, narrow evaluation reorganizes capability concepts: architectures and definitions are selected for benchmark legibility until evaluation ceases to track an independent object and instead produces a version of the target defined by its own operational assumptions. The result is a trap: evaluation frameworks treat self-reinforcing assessments as valid, both creating and obscuring structural limits on what the current paradigm can accomplish. We introduce Epistematics, a methodology for deriving evaluation criteria directly from technical capability claims and auditing whether proposed benchmarks can discriminate the claimed capability from proxy behaviors. The contribution is meta-evaluative: an audit procedure, a failure mode taxonomy, and benchmark-design criteria for evaluating capability-evaluation coherence. We demonstrate the procedure through a worked audit of Dupoux et al. (2026), a proposal that revises the dominant paradigm's theoretical assumptions at the architectural level while reproducing them in its evaluation criteria, thereby entrenching the constraint it seeks to overcome in a form the evaluation cannot detect.

Background: Many different approaches to controlled text generation (CTG) have been proposed over recent years, but it is difficult to get a clear picture of which approach performs best, because different datasets and evaluation methods are used in each case to assess the control achieved. Objectives: Our aim in the work reported in this paper is to develop an approach to evaluation that enables us to comparatively evaluate different CTG systems in a manner that is both informative and fair to the individual systems. Methods: We use a level-playing-field (LPF) approach to comparative evaluation where we (i) generate and process all system outputs in a standardised way, and (ii) apply a shared set of evaluation methods and datasets, selected based on those currently in use, in order to ensure fair evaluation. Results: When re-evaluated in this way, performance results for a representative set of current CTG systems differ substantially from originally reported results, in most cases for the worse. This highlights the importance of a shared standardised way of assessing controlled generation. Conclusions: The discrepancies revealed by LPF evaluation demonstrate the urgent need for standardised, reproducible evaluation practices in CTG. Our results suggest that without such practices, published performance claims may substantially misrepresent true system capabilities.

Recent advances in de novo protein binder design have enabled increasing experimental validation, yet reported in silico metrics remain difficult to interpret or compare across studies due to non-standardized evaluation protocols. We introduce ProtDBench, a standardized and throughput-aware evaluation framework for protein binder design. ProtDBench defines unified benchmark tasks, evaluation protocols, and success criteria, enabling systematic analysis of how evaluation design influences observed performance. Using a large wet-lab annotated dataset, we analyze commonly used structure prediction models as evaluation verifiers, revealing substantial verifier-dependent bias and limited agreement under identical filtering protocols. We then benchmark representative open-source generative binder design methods across ten diverse protein targets under a fixed evaluation protocol. Beyond per-sequence success rates, ProtDBench incorporates throughput-aware metrics based on a fixed 24-hour budget, as well as cluster-level success criteria to account for structural diversity. Together, these results expose systematic differences induced by filtering rules, success definitions, and throughput-aware evaluation between computational efficiency, success rate, and structural diversity. Overall, ProtDBench provides a fair and reproducible evaluation pipeline that supports systematic and controlled comparison of protein binder design methods under realistic evaluation settings.

Evaluating LLM-generated business ideas is often harder to scale than generating them. Unlike standard NLP benchmarks, business idea evaluation relies on multi-dimensional criteria such as feasibility, novelty, differentiation, user need, and market size, and expert judgments often disagree. This paper studies a methodological question raised by such disagreement: should an automatic judge approximate an aggregate consensus, or model evaluators individually? We introduce PBIG-DATA, a dataset of approximately 3,000 individual scores across 300 patent-grounded product ideas, provided by domain experts on six business-oriented dimensions: specificity, technical validity, innovativeness, competitive advantage, need validity, and market size. Analyses show substantial expert disagreement on fine-grained ordinal scores, while agreement is higher under coarse selection, suggesting structured heterogeneity rather than random noise. We then compare three judge configurations: a rubric-only zero-shot judge, an aggregate judge conditioned on mixed evaluator histories, and a personalized judge conditioned on the target evaluator's scoring history. Across dimensions and model sizes, personalized judges align more closely with the corresponding evaluator than aggregate judges, and evaluator agreement correlates with similarity of judge-generated reasoning only under personalized conditioning. These results indicate that pooled labels can be a fragile target in pluralistic evaluation settings and motivate evaluator-conditioned judge designs for business idea assessment.

Large Vision-Language Models (VLMs) are increasingly used to evaluate outputs of other models, for image-to-text (I2T) tasks such as visual question answering, and text-to-image (T2I) generation tasks. Despite this growing reliance, the reliability of these Evaluator VLMs remains under explored. In this work, we systematically evaluate the reliability of Evaluator VLMs across both I2T and T2I tasks. We introduce targeted perturbations that degrade output quality along key error dimensions, including object hallucinations, spatial reasoning, factual grounding, and visual fidelity. These perturbations test whether Evaluator VLMs can reliably account for these quality degrading errors in their evaluations. Using a comprehensive benchmark of over 4000 perturbed instances spanning 40 perturbation dimensions, we evaluate 4 prominent VLMs using single-answer scoring, pairwise comparison, and reference-guided paradigms. Our findings reveal that current VLM evaluators exhibit substantial blind spots: they often fail to detect perturbed outputs - in some cases exceeding 50%, struggle particularly with fine-grained compositional and spatial errors, and are often insensitive to hallucinated content that contradicts the input image. Pairwise comparison proves more reliable, though failure rates persist. These results highlight the unreliable nature of current Evaluator VLMs and urge caution in their deployment for benchmarking and development decisions. Code and data have been made publicly available.

Federated pre-training offers a way to train foundation models on private or distributed data without centralizing the underlying datasets. However, evaluating federated pre-training remains challenging because differences in client participation and local data availability can make directly comparable evaluation difficult. Moreover, pre-training test perplexity is tied to the pre-training distribution, while downstream benchmarks introduce task-specific adaptation that may not faithfully reflect the test perplexity established during pre-training. In this work, we study which evaluation protocol more reliably reflects federated pre-training quality. Using a controlled set of centralized and federated-trained models of a 16M parameter transformer model trained on identical client data, we assess evaluation protocols by whether they preserve a reference ranking established on the same pre-training testset. We compare downstream fine-tuning on GLUE, including full, head-only, and reduced-data variants, with next-token prediction on GLUE text as an intrinsic evaluation signal. Our results show that downstream fine-tuning does not reliably preserve the pre-training ranking, whereas direct next-token prediction exhibits a strong correspondence with the pre-training test perplexity. These findings suggest that downstream fine-tuning alone can be misleading when comparing federated pre-trained models, and that evaluation signals closer to the original pre-training objective deserve greater attention.

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.

As Generative AI (GenAI) tools become increasingly capable of generating solutions to computing assignments, the computing education community is exploring pedagogical approaches that emphasize solution evaluation, verification, and critique alongside traditional solution generation. However, evidence regarding the impact of such evaluation-centered tasks on student learning remains limited, particularly in upper-division, theory-heavy courses. We conducted a randomized A/B crossover study (N=220) in a junior-level algorithms course to compare evaluating GenAI-generated solutions with traditional problem solving. Across six assignments, student working groups either solved challenging algorithmic problems directly or evaluated often-flawed GenAI-generated solutions, with roles reversed midway through the semester. We found no statistically significant differences between groups in midterm scores, final exam scores, overall course grades, or exam problems structurally aligned with the homework interventions. Students received significantly higher homework scores when evaluating GenAI-generated solutions, but this localized advantage did not translate into downstream summative gains. Survey data further indicated that most students reported no change in study habits in response to the intervention; however, those who reported adapting their study strategies rated the GenAI-evaluation assignments as significantly more helpful. These findings suggest that GenAI evaluation redistributes student effort from open-ended solution construction toward verification, diagnosis, and judgment, but does not automatically produce stronger conceptual transfer. We conclude that GenAI-evaluation activities can be incorporated into algorithms coursework without broad performance losses, but meaningful learning gains may require deliberate scaffolding that pushes students beyond simple error diagnosis.

In measurement theory, instruments do not simply record reality; they help constitute what is observed. The same holds for generative AI evaluation: benchmarks do not just measure, they shape what models appear to be. Functionalist benchmarks treat models as isolated predictors, while prescriptive approaches assess what systems ought to be. Both obscure the sociotechnical processes through which meaning and values are enacted, risking the reification of narrow cultural perspectives in pluralist contexts. This thesis advances a descriptive alternative. It argues that generative AI must be evaluated as a pluralist sociotechnical system and develops Machine-Society-Human (MaSH) Loops, a framework for tracing how models, users, and institutions recursively co-construct meaning and values. Evaluation shifts from judging outputs to examining how values are enacted in interaction. Three contributions follow. Conceptually, MaSH Loops reframes evaluation as recursive, enactive process. Methodologically, the World Values Benchmark introduces a distributional approach grounded in World Values Survey data, structured prompt sets, and anchor-aware scoring. Empirically, the thesis demonstrates these through two cases: value drift in early GPT-3 and sociotechnical evaluation in real estate. A final chapter draws on participatory realism to argue that prompting and evaluation are constitutive interventions, not neutral observations. The thesis argues that static benchmarks are insufficient for generative AI. Responsible evaluation requires pluralist, process-oriented frameworks that make visible whose values are enacted. Evaluation is therefore a site of governance, shaping how AI systems are understood, deployed, and trusted.

Modern AI agents increasingly rely on search infrastructure to execute complex, neuro-symbolic reasoning workflows. These workflows often compile into deeply nested, non-monotonic Boolean queries over text fields. However, standard query evaluation strategies over inverted indices face severe theoretical limits when handling these structures. Stateful iterator models (Document-at-a-Time) are structurally bounded by $\text{NC}^1$ formula evaluation, suffering a worst-case $O(2^{|Q|})$ exponential blowup in query complexity when unrolling re-convergent logic. Conversely, recursive materialization models (Term-at-a-Time) incur an $Ω(|U|)$ space complexity penalty (the Universal Scan) when evaluating logical negation over the document universe. In this paper, we establish the theoretical boundaries of executing complex logic natively over an inverted index. We formalize a retrieval language ($\mathcal{L}_R$) based on Directed Acyclic Graphs (DAGs) and prove that its evaluation problem is strictly \textbf{$\mathbf{P}$-Complete}. To make evaluation tractable, we introduce \texttt{ComputePN}, a deterministic, sparsity-aware evaluation algorithm. By decoupling logical negation from universe-scale materialization via a novel Positive-Negative dual representation, and utilizing native DAG memoization, \texttt{ComputePN} strictly bounds evaluation time to $O(|Q| \cdot |U_{\mathit{active}}|)$. This approach successfully evaluates $\mathbf{P}$-Complete queries natively over the index, avoiding both the combinatorial tree-expansion bottleneck and the universal scan penalty, laying the formal foundation for computational retrieval.

Autonomous driving has shifted towards end-to-end policy learning, where reliable, interpretable policy evaluation is a fundamental challenge as driving quality is highly context-dependent. Commonly used rule-based driving metrics like EPDMS are interpretable but lack context-awareness, while recent VLMbased evaluations are context-aware but limited by ambiguous VLM outputs and weak physical grounding. To evaluate driving in a manner that is both interpretable and context-aware, we introduce DriveJudge. DriveJudge is a driving evaluation agent that combines rule-grounded evaluation with Vision-Language Model (VLM) reasoning and selectively invokes physically-grounded deterministic rule functions after interpreting the environmental context. To train and evaluate DriveJudge, we curate a large-scale dataset of 33,577 challenging driving samples with human annotations on whether the driving behavior is reasonable in the given scenario. With this dataset, we address the underexplored problem of driving metric evaluation, and introduce two human-aligned benchmark tasks: Driving Quality Classification and Trajectory Preference Selection. DriveJudge outperforms EPDMS for driving quality classification by 21.23 AUC, and the recent VLM-based DriveCritic for trajectory preference selection by 6.5%, setting a new standard for interpretable and precise driving evaluation.

Knowledge graph completion (KGC) models are commonly evaluated using rank-based metrics such as MRR and Hits@K, despite different users often requiring different evaluation perspectives. In this demo, we present PROBE-Web, an interactive system for probing diverse evaluation landscapes for KGC models. PROBE-Web enables users to flexibly evaluate KGC models by adjusting two critical perspectives: (P1) predictive sharpness and (P2) popularity-bias robustness. Through a user-friendly GUI, users easily evaluate multiple KGC models and analyze their strengths and weaknesses. PROBE-Web provides four key functionalities: (1) conventional evaluation toolkit, (2) flexible perspective-aware evaluation, (3) explainable case studies, and (4) evaluation landscape exploration. We believe that PROBE-Web can help users better understand KGC models aligning with their objectives.

Frontier language models sometimes recognize that they are being evaluated and adjust their behavior, undermining validity of benchmark results. Yet the field studies it without a shared foundation, conflating properties of the evaluation with properties of the model, and detection with behavioral response. We ground evaluation awareness in social psychology, decomposing it into an environment component (how recognizable the task is) and a model component that separates recognition from propensity to act on it. We operationalize the environment component through eight categorized trigger factors, such as placeholder entities and grading-style output formats, and study recognition and behavior through chain-of-thought monitoring. Across nine frontier models and four benchmarks, recognition rates depend on the specific pairing of model and benchmark rather than on either in isolation. Recognition rarely leads to behavioral change, and when it does, the direction depends on the type of evaluation perceived. Models are also more sensitive to safety than capability evaluations, placing safety benchmark validity at greater risk. To study which factors each model is sensitive to and how they interact, we propose \textbf{EvalAwareBench}, a factor-controlled benchmark of 100 paired safety-capability tasks where each of the eight factors can be independently toggled, varying evaluative signals while holding the underlying request fixed. Through EvalAwareBench, we find that no single factor uniformly affects all models, but stacking factors progressively raises evaluation awareness across all of them. Our framework and EvalAwareBench provide the tools to measure, attribute, and mitigate evaluation awareness, pointing to behavioral consistency under recognition as a promising path forward.

In this study, we propose a framework that incorporates subjective evaluations provided by a Vision-Language Model (VLM) into the fitness evaluation and selection processes of a genetic algorithm. As the target of evolution, we employ virtual soft robots with flexible morphologies and locomotion and present the VLM with sequence images representing the locomotion of two individuals. Selection is performed via pairwise comparisons based on subjective evaluation terms such as adorably and weirdly. The outcomes of these comparisons are used as selection pressure within the genetic algorithm, enabling the simultaneous evolution of morphology and locomotion. Experimental results demonstrate that subjective selection by the VLM accelerates population convergence compared to random selection, while also giving rise to distinctive morphologies and motions corresponding to each evaluation term. An auxiliary experiment with human participants further showed that, although individual pairwise choices only partly agreed with the VLM selections, the resulting morphological and locomotion tendencies were qualitatively similar and repeated human evaluations imposed noticeable fatigue. Moreover, the observation that similar evolutionary outcomes emerged across different evaluation terms suggests that the VLM does not apply these terms in a purely literal manner but instead decomposes them into multiple internal evaluation criteria when making judgments. This work visualizes the evolutionary process through which subjective linguistic expressions are mapped onto embodied phenotypes and provides a foundational framework for analyzing the structure of subjective judgment in VLMs. The proposed approach is expected to contribute to new developments in evolutionary computation and artificial life research based on subjective evaluation.

Generalist robot manipulation policies trained on large, diverse datasets have shown remarkable promise across a wide range of tasks. However, rigorously evaluating these policies remains a fundamental challenge. Real-world performance depends on a large combinatorial space of task factors including object poses and camera viewpoints, making full, exhaustive evaluation intractable. Additionally, real hardware evaluation is slow and resource-intensive, so current practice is to use narrow test suites that can miss critical failure modes and misrepresent true deployment readiness. We propose an active evaluation framework that addresses this challenge by treating policy evaluation as a sequential experimental design problem. Our approach fits a probabilistic surrogate model over a structured space of task factors and adaptively selects evaluation configurations to maximize information gain over the policy's performance distribution, allowing for sample-efficient characterization of policy behavior across unseen conditions and a systematic identification of failure-prone regions. We conduct 2331 real-world evaluations across 3 tasks with 3 factor variations and find that our approach typically saves the evaluator at least 20-40% of trials compared to typical random testing.

Evaluating generative AI models is a routine, but resource-intensive, process that is conducted over and over again during the course of model development. In this work, we propose Collaborative Evaluation (CollabEval), a simple, effective, and principled method for exploiting dependencies between historical runs of different models on the same tasks to improve statistical efficiency. Specifically, our approach treats model evaluation as a matrix completion problem over an $M \times N$ matrix of evaluation scores, where $M$ is the total number of models and $N$ is the total number of evaluation prompts. We assume that a subset of these $M$ models are targeted for evaluation. For these target models only a small fraction, $p$, of prompts has been annotated with evaluation scores. Leveraging recent results in prediction-powered inference, we build a low-rank approximation of the score matrix, and use the reconstructed values as control variates in a manner that guarantees unbiased estimates of the true evaluation metric mean, in addition to statistically valid confidence intervals. Empirically, across a wide range of datasets, models, and sparsity levels $p$, we find that CollabEval substantially reduces the mean confidence interval size, and the mean squared error of the point estimate, compared to baseline methods at the same annotation budget.

Evaluating embodied robot foundation models remains a critical bottleneck; unlike large language models efficiently assessed via digital benchmarks, robotic policies require slow, costly real-world rollouts limited by hardware and human supervision, which has driven interest in world models as surrogate policy evaluators, yet the key properties that make a world model reliable for policy assessment remain poorly understood. This work presents a systematic study of world models for robotic policy evaluation and introduces WMBench, a benchmark constructed from real-robot teleoperation data and matched policy rollouts covering diverse manipulation tasks to enable controlled comparisons across model families, action encodings, rollout horizons, and evaluation metrics. Using WMBench, we analyze 7 video world models, 4 action representation schemes, and over 324,000 simulated policy rollouts paired with real robot executions, further enriching our analysis with large-scale community submissions from the CVPR 2026 GigaBrain Challenge, curated synthetic trajectories, and a training videos spanning more than 12,000 hours. Our experiments deliver three core insights: evaluator quality is dominated by long-horizon, action-faithful rollout consistency rather than short-term visual realism; pretraining gains stem not only from data scale but from balancing general world knowledge with robot-specific controllability; and architectural choices including action encoding, memory design, and evaluator-focused post-training strongly determine alignment with real-world robot behavior. Drawing on these results, we derive a practical design roadmap and realize it in \textit{GigaWorld-1}, a world model specially optimized for policy evaluation, and we fully release our code, models, datasets, and toolkits to advance scalable evaluation research for embodied foundation models.

Evaluation in machine learning is typically treated as a neutral measurement process. However, in operational information systems, evaluation outcomes are often conditioned by the processes used to generate labels. This paper does not seek to improve classification performance. Instead, it examines the validity of performance measurement under differing label-authority regimes. This issue is particularly relevant in large-scale metadata-driven systems, where labels are often incomplete, inconsistent, or weakly supervised. We introduce evaluation sovereignty, defined as the degree to which performance metrics are independent of label authority and supervision regime, and propose a multi-track evaluation framework that systematically varies training and evaluation label sources. Using hierarchical multi-label classification on large-scale scientific metadata, we demonstrate that models exhibiting strong performance under operational ("silver") evaluation degrade substantially under independent ("gold") evaluation, particularly for fine-grained classification. For example, Micro-F1 decreases from approximately 0.54 to 0.03. Notably, ranking-based metrics remain above baseline, revealing a divergence between latent model signal and classification validity. These findings suggest that commonly reported performance metrics may reflect alignment with labeling processes rather than true predictive capability. We therefore reconceptualize evaluation validity as a system-level property shaped by label governance and provide a practical methodology for auditing intelligent systems operating under weak supervision.

Evaluation remains a critical bottleneck for interactive agent development. Existing evaluation methods often rely on static benchmarks, which fail to capture the dynamic, multi-step nature of agentic behavior and struggle to expose meaningful failure modes. While user-simulation-based evaluation offers a promising alternative, existing simulation frameworks suffer from two major limitations. First, they provide limited mechanisms for evaluating the quality and comprehensiveness of simulated interactions, making it difficult to assess whether a simulator sufficiently explores an agent's capabilities and failure modes. Second, most frameworks are restricted to either UI-only actions or API-only actions, limiting their ability to model the full range of realistic user behaviors. To address these limitations, we propose VISTA, a Versatile Interactive user Simulation Toolkit for Agent evaluation. Our toolkit includes a suite of six metrics for measuring the realism, capability coverage, and interaction effectiveness of simulated interactions. In addition, we develop a hybrid user simulator that integrates both UI-based interactions and API-based interactions, enabling more realistic and comprehensive evaluation across diverse interactive environments. We evaluate VISTA in e-commerce shopping and education customer service settings and demonstrate that it produces more realistic and comprehensive evaluations than existing methods.

As large language models (LLMs) are increasingly used for long-form generation, reliably evaluating long-form outputs has become a critical challenge. LLM-as-a-judge offers a scalable alternative to human evaluation, yet its reliability in long-form output evaluation remains underexamined: existing meta-evaluation benchmarks focus mainly on short-form outputs. Compared with short-form evaluation, long-form evaluation is not merely a matter of output length; it often requires judges to make more complex document-level assessments of overall organization, task-relevant coverage and depth, cross-section consistency, and scenario-specific quality criteria. In this work, we introduce LongJudgeBench, a comprehensive benchmark for evaluating LLM judges on long-form outputs across diverse real-world scenarios and judging protocols. We systematically evaluate a broad range of LLM judges, covering multiple base models and judging settings. Our results reveal a substantial reliability gap: current LLM judges remain unstable across scenarios, and rubrics or references are helpful but not always sufficient. We hope LongJudgeBench will support future research on more robust, context-aware, and human-aligned LLM-as-a-judge methods. Our code is available at https://github.com/cjj826/LongJudgeBench.

As Large Language Models (LLMs) evolve from general-purpose assistants to user-centric agents, personalization has become central to aligning model behavior with individual preferences, making the evaluation of personalized alignment a critical bottleneck. Existing evaluation methods-ranging from automatic metrics to LLM-as-a-judge approaches-fail to capture subjective, user-specific preferences embedded in long-term interaction histories. We identify three essential principles for reliable and effective personalized evaluation: Representativeness, User-Consistency, and Discriminativeness. To address these principles, we introduce Personalized Evaluation as Learning, a paradigm that formulates personalized evaluation as a learning problem rather than a static judgment. Under this paradigm, we propose PARL (Preference-Aware Rubric Learning for Personalized Evaluation), a framework that learns to induce preference-aware evaluation rubrics directly from raw user histories and performs a self-validation mechanism to ensure consistency with the user's preferences. PARL integrates rubric induction with a discriminative reinforcement learning objective that contrasts user-authored responses against competitive personalized model outputs, enabling the learned rubrics to capture precise, user-specific decision boundaries. Experiments on real-world personalized text generation tasks show that PARL consistently induces high-fidelity rubrics that reliably identify user-aligned responses and generalize across users and tasks, while capturing stable stylistic preferences and fine-grained evaluative patterns. To ensure reproducibility, our code is available at https://github.com/SnowCharmQ/PARL.

AI agents are entering high-risk production settings, where they use tools, retain context, follow policies, handle private data, and interact with users over multiple turns. Yet many evaluation methods still judge isolated outputs or static tasks, missing failures that emerge through trajectory, pressure, and adversarial interaction. We introduce ProofAgent Harness, open infrastructure for scalable, auditable, and adversarial AI agent evaluation. The harness provides evaluation infrastructure around an agent: it curates evaluation intelligence, runs adversarial multi-turn trials, captures behavioral traces, applies post-hoc multi-juror scoring, resolves disagreement, and produces evidence-linked reports. Its open design allows developers and researchers to extend domains, traps, metrics, juror personas, scoring rules, and reporting formats. At its core is Adversarial Multi-Juror Scoring with Turn-Level Audit, which evaluates completed agent behavior under pressure using calibrated juror personas, consensus checks, and turn-level evidence. Experiments across customer support, medical triage, privacy and security, and code generation agents show that strong agents fail selectively through weak metrics, fragile turns, unsafe reframing, and manipulation paths. We also find that a small quantized local Harness LLM can challenge production agents powered by best-in-class large LLMs, suggesting that evaluation capability emerges from the full harness pipeline rather than model scale alone. ProofAgent Harness turns AI agent evaluation from a static score into scalable adversarial evaluation infrastructure: repeatable, evidence-backed, extensible, and actionable before deployment.

Evaluating large language models (LLMs) for clinical documentation tasks such as SOAP note generation remains challenging. Unlike standard summarization, these tasks require clinical abstraction, normalization of colloquial language, and medically grounded inference. However, prevailing evaluation methods including automated metrics and LLM as judge frameworks rely on lexical faithfulness, often labeling any information not explicitly present in the transcript as hallucination. We show that such approaches systematically misclassify clinically valid outputs as errors, inflating hallucination rates and distorting model assessment. Our analysis reveals that many flagged hallucinations correspond to legitimate clinical transformations, including synonym mapping, abstraction of examination findings, diagnostic inference, and guideline consistent care planning. By aligning evaluation criteria with clinical reasoning through calibrated prompting and retrieval grounded in medical ontologies we observe a significant shift in outcomes. Under a lexical evaluation regime, the mean hallucination rate is 35%, heavily penalizing valid reasoning. With inference aware evaluation, this drops to 9%, with remaining cases reflecting genuine safety concerns. These findings suggest that current evaluation practices over penalize valid clinical reasoning and may measure artifacts of evaluation design rather than true errors, underscoring the need for clinically informed evaluation in high context domains like medicine.

Vision-Language-Action (VLA) models are increasingly evaluated across multiple simulation benchmarks, yet adding each benchmark to an evaluation pipeline requires resolving incompatible dependencies, matching underspecified evaluation protocols, and reverse-engineering undocumented preprocessing. This burden scales with the number of models and benchmarks, making comprehensive evaluation impractical for most teams. We present vla-eval, an open-source evaluation harness that eliminates this per-benchmark cost by decoupling model inference from benchmark execution through a WebSocket+msgpack protocol with Docker-based environment isolation. Models integrate once by implementing a single predict() method; benchmarks integrate once via a four-method interface; the full cross-evaluation matrix works automatically. The framework supports 14 simulation benchmarks and six model servers. Parallel evaluation via episode sharding and batch inference achieves up to 47x wall-clock speedup, completing 2,000 LIBERO episodes in ~18 minutes. To validate the framework, we reproduce published scores across six VLA codebases and three benchmarks, documenting previously undocumented pitfalls. We additionally release a VLA leaderboard aggregating 657 published results across 17 benchmarks. Framework, evaluation configs, and all reproduction results are publicly available at https://github.com/allenai/vla-evaluation-harness and https://allenai.github.io/vla-evaluation-harness/leaderboard.