Daily Papers

Machine unlearning is a process to remove specific data points from a trained model while maintaining the performance on retain data, addressing privacy or legal requirements. Despite its importance, existing unlearning evaluations tend to focus on logit-based metrics (i.e., accuracy) under small-scale scenarios. We observe that this could lead to a false sense of security in unlearning approaches under real-world scenarios. In this paper, we conduct a new comprehensive evaluation that employs representation-based evaluations of the unlearned model under large-scale scenarios to verify whether the unlearning approaches genuinely eliminate the targeted forget data from the model's representation perspective. Our analysis reveals that current state-of-the-art unlearning approaches either completely degrade the representational quality of the unlearned model or merely modify the classifier (i.e., the last layer), thereby achieving superior logit-based evaluation metrics while maintaining significant representational similarity to the original model. Furthermore, we introduce a rigorous unlearning evaluation setup, in which the forgetting classes exhibit semantic similarity to downstream task classes, necessitating that feature representations diverge significantly from those of the original model, thus enabling a more rigorous evaluation from a representation perspective. We hope our benchmark serves as a standardized protocol for evaluating unlearning algorithms under realistic conditions.

The automatic evaluation of instruction following typically involves using large language models (LLMs) to assess response quality. However, there is a lack of comprehensive evaluation of these LLM-based evaluators across two dimensions: the base LLMs and the evaluation protocols. Therefore, we present a thorough meta-evaluation of instruction following, including 25 base LLMs and 15 recently proposed evaluation protocols, on 4 human-annotated datasets, assessing the evaluation accuracy of the LLM-evaluators. Our evaluation allows us to identify the best-performing base LLMs and evaluation protocols with a high degree of robustness. Moreover, our large-scale evaluation reveals: (1) Base LLM performance ranking remains largely consistent across evaluation protocols, with less capable LLMs showing greater improvement from protocol enhancements; (2) Robust evaluation of evaluation protocols requires many base LLMs with varying capability levels, as protocol effectiveness can depend on the base LLM used; (3) Evaluation results on different datasets are not always consistent, so a rigorous evaluation requires multiple datasets with distinctive features. We release our meta-evaluation suite ReIFE, which provides the codebase and evaluation result collection for more than 500 LLM-evaluator configurations, to support future research in instruction-following evaluation.

The current state of machine learning scholarship in Timeseries Anomaly Detection (TAD) is plagued by the persistent use of flawed evaluation metrics, inconsistent benchmarking practices, and a lack of proper justification for the choices made in novel deep learning-based model designs. Our paper presents a critical analysis of the status quo in TAD, revealing the misleading track of current research and highlighting problematic methods, and evaluation practices. Our position advocates for a shift in focus from solely pursuing novel model designs to improving benchmarking practices, creating non-trivial datasets, and critically evaluating the utility of complex methods against simpler baselines. Our findings demonstrate the need for rigorous evaluation protocols, the creation of simple baselines, and the revelation that state-of-the-art deep anomaly detection models effectively learn linear mappings. These findings suggest the need for more exploration and development of simple and interpretable TAD methods. The increment of model complexity in the state-of-the-art deep-learning based models unfortunately offers very little improvement. We offer insights and suggestions for the field to move forward. Code: https://github.com/ssarfraz/QuoVadisTAD

While machine learning on graphs has demonstrated promise in drug design and molecular property prediction, significant benchmarking challenges hinder its further progress and relevance. Current benchmarking practices often lack focus on transformative, real-world applications, favoring narrow domains like two-dimensional molecular graphs over broader, impactful areas such as combinatorial optimization, relational databases, or chip design. Additionally, many benchmark datasets poorly represent the underlying data, leading to inadequate abstractions and misaligned use cases. Fragmented evaluations and an excessive focus on accuracy further exacerbate these issues, incentivizing overfitting rather than fostering generalizable insights. These limitations have prevented the development of truly useful graph foundation models. This position paper calls for a paradigm shift toward more meaningful benchmarks, rigorous evaluation protocols, and stronger collaboration with domain experts to drive impactful and reliable advances in graph learning research, unlocking the potential of graph learning.

Automated fine-grained perception of calligraphy styles--a task vital to cultural heritage preservation--remains a critical challenge for Large Vision-Language Models (LVLMs), largely constrained by existing datasets that suffer from modal mixture and flattened labels. To bridge this gap, we introduce HCSU, the first comprehensive dataset tailored for fine-grained Historical Calligraphy Style Understanding. HCSU comprises 39,307 meticulously curated character images from 49 historically prominent calligraphers across 10 dynasties, systematically decoupling authentic ink manuscripts (Tie) from stone rubbings (Bei) to resolve the long-standing modal mixture problem. Moving beyond conventional flattened labels, HCSU provides hierarchical expert-written aesthetic descriptions, enabling two rigorous evaluation protocols: fine-grained style discrimination and interpretable aesthetic reasoning. Extensive evaluations reveal a persistent gap between calligraphy-related knowledge and visually grounded style perception: state-of-the-art LVLMs show non-trivial performance but remain sensitive to script-level, textual, and source-specific cues, and often struggle to ground aesthetic judgments in fine-grained brushwork evidence. Ultimately, the HCSU benchmark exposes fundamental limitations in current multimodal architectures, aiming to inspire the evolution of expert-level visual reasoning for cultural heritage preservation. The dataset is available at https://huggingface.co/datasets/Tongji209/HCSU.

Nuclear Magnetic Resonance (NMR) spectroscopy is the cornerstone of small-molecule structure elucidation. While deep learning has demonstrated significant potential in automating structure elucidation and spectral simulation, current progress is severely impeded by the reliance on synthetic datasets, which introduces significant domain shifts when applied to real-world experimental spectra. Furthermore, the lack of standardized evaluation protocols and rigorous data splitting strategies frequently leads to unfair comparisons and data leakage. To address these challenges, we introduce NMRGym, the largest and most comprehensive standardized dataset and benchmark derived from high-quality experimental NMR data to date. Comprising 269,999 unique molecules paired with high-fidelity ^1H and ^{13}C spectra, NMRGym bridges the critical gap between synthetic approximations and real-world diversity. We implement a strict quality control pipeline and unify data formats to ensure fair comparison. To strictly prevent data leakage, we enforce a scaffold-based split. Additionally, we provide fine-grained peak-atom level annotations to support future usage. Leveraging this resource, we establish a comprehensive evaluation suite covering diverse downstream tasks, including structure elucidation, functional group prediction from NMR, toxicity prediction from NMR, and spectral simulation, benchmarking representative state-of-the-art methodologies. Finally, we release an open-source leadboard with an automated leaderboard to foster community collaboration and standardize future research. The dataset, benchmark and leaderboard are publicly available at blue{https://AIMS-Lab-HKUSTGZ.github.io/NMRGym/}.

3D Gaussian Splatting (3DGS) has revolutionized fast novel view synthesis, yet its opacity-based formulation makes surface extraction fundamentally difficult. Unlike implicit methods built on Signed Distance Fields or occupancy, 3DGS lacks a global geometric field, forcing existing approaches to resort to heuristics such as TSDF fusion of blended depth maps. Inspired by the Objects as Volumes framework, we derive a principled occupancy field for Gaussian Splatting and show how it can be used to extract highly accurate watertight meshes of complex scenes. Our key contribution is to introduce a learnable oriented normal at each Gaussian element and to define an adapted attenuation formulation, which leads to closed-form expressions for both the normal and occupancy fields at arbitrary locations in space. We further introduce a novel consistency loss and a dedicated densification strategy to enforce Gaussians to wrap the entire surface by closing geometric holes, ensuring a complete shell of oriented primitives. We modify the differentiable rasterizer to output depth as an isosurface of our continuous model, and introduce Primal Adaptive Meshing for Region-of-Interest meshing at arbitrary resolution. We additionally expose fundamental biases in standard surface evaluation protocols and propose two more rigorous alternatives. Overall, our method Gaussian Wrapping sets a new state-of-the-art on DTU and Tanks and Temples, producing complete, watertight meshes at a fraction of the size of concurrent work-recovering thin structures such as the notoriously elusive bicycle spokes.

Given a large and evolving codebase, the ability to automatically generate holistic, architecture-aware documentation that captures not only individual functions but also cross-file, cross-module, and system-level interactions remains an open challenge. Comprehensive documentation is essential for long-term software maintenance and collaboration, yet current automated approaches still fail to model the rich semantic dependencies and architectural structures that define real-world software systems. We present CodeWiki, a unified framework for automated repository-level documentation across seven programming languages. CodeWiki introduces three key innovations: (i) hierarchical decomposition that preserves architectural context across multiple levels of granularity, (ii) recursive multi-agent processing with dynamic task delegation for scalable generation, and (iii) multi-modal synthesis that integrates textual descriptions with visual artifacts such as architecture diagrams and data-flow representations. To enable rigorous evaluation, we introduce CodeWikiBench, a comprehensive benchmark featuring multi-dimensional rubrics and LLM-based assessment protocols. Experimental results show that CodeWiki achieves a 68.79\% quality score with proprietary models, outperforming the closed-source DeepWiki baseline (64.06\%) by 4.73\%, with particularly strong improvements on high-level scripting languages (+10.47\%). We open-source CodeWiki to foster future research and community adoption.

While deep learning has revolutionized computer-aided drug discovery, the AI community has predominantly focused on model innovation and placed less emphasis on establishing best benchmarking practices. We posit that without a sound model evaluation framework, the AI community's efforts cannot reach their full potential, thereby slowing the progress and transfer of innovation into real-world drug discovery. Thus, in this paper, we seek to establish a new gold standard for small molecule drug discovery benchmarking, WelQrate. Specifically, our contributions are threefold: WelQrate Dataset Collection - we introduce a meticulously curated collection of 9 datasets spanning 5 therapeutic target classes. Our hierarchical curation pipelines, designed by drug discovery experts, go beyond the primary high-throughput screen by leveraging additional confirmatory and counter screens along with rigorous domain-driven preprocessing, such as Pan-Assay Interference Compounds (PAINS) filtering, to ensure the high-quality data in the datasets; WelQrate Evaluation Framework - we propose a standardized model evaluation framework considering high-quality datasets, featurization, 3D conformation generation, evaluation metrics, and data splits, which provides a reliable benchmarking for drug discovery experts conducting real-world virtual screening; Benchmarking - we evaluate model performance through various research questions using the WelQrate dataset collection, exploring the effects of different models, dataset quality, featurization methods, and data splitting strategies on the results. In summary, we recommend adopting our proposed WelQrate as the gold standard in small molecule drug discovery benchmarking. The WelQrate dataset collection, along with the curation codes, and experimental scripts are all publicly available at WelQrate.org.

Integrating Large Language Models with symbolic planners is a promising direction for obtaining verifiable and grounded plans compared to planning in natural language, with recent works extending this idea to visual domains using Vision-Language Models (VLMs). However, rigorous comparison between VLM-grounded symbolic approaches and methods that plan directly with a VLM has been hindered by a lack of common environments, evaluation protocols and model coverage. We introduce ViPlan, the first open-source benchmark for Visual Planning with symbolic predicates and VLMs. ViPlan features a series of increasingly challenging tasks in two domains: a visual variant of the classic Blocksworld planning problem and a simulated household robotics environment. We benchmark nine open-source VLM families across multiple sizes, along with selected closed models, evaluating both VLM-grounded symbolic planning and using the models directly to propose actions. We find symbolic planning to outperform direct VLM planning in Blocksworld, where accurate image grounding is crucial, whereas the opposite is true in the household robotics tasks, where commonsense knowledge and the ability to recover from errors are beneficial. Finally, we show that across most models and methods, there is no significant benefit to using Chain-of-Thought prompting, suggesting that current VLMs still struggle with visual reasoning.

A comprehensive benchmark is yet to be established in the Image Manipulation Detection & Localization (IMDL) field. The absence of such a benchmark leads to insufficient and misleading model evaluations, severely undermining the development of this field. However, the scarcity of open-sourced baseline models and inconsistent training and evaluation protocols make conducting rigorous experiments and faithful comparisons among IMDL models challenging. To address these challenges, we introduce IMDL-BenCo, the first comprehensive IMDL benchmark and modular codebase. IMDL-BenCo: i) decomposes the IMDL framework into standardized, reusable components and revises the model construction pipeline, improving coding efficiency and customization flexibility; ii) fully implements or incorporates training code for state-of-the-art models to establish a comprehensive IMDL benchmark; and iii) conducts deep analysis based on the established benchmark and codebase, offering new insights into IMDL model architecture, dataset characteristics, and evaluation standards. Specifically, IMDL-BenCo includes common processing algorithms, 8 state-of-the-art IMDL models (1 of which are reproduced from scratch), 2 sets of standard training and evaluation protocols, 15 GPU-accelerated evaluation metrics, and 3 kinds of robustness evaluation. This benchmark and codebase represent a significant leap forward in calibrating the current progress in the IMDL field and inspiring future breakthroughs. Code is available at: https://github.com/scu-zjz/IMDLBenCo.

While Diffusion Large Language Models (DLLMs) have demonstrated remarkable capabilities in multi-modal generation, performing precise, training-free image editing remains an open challenge. Unlike continuous diffusion models, the discrete tokenization inherent in DLLMs hinders the application of standard noise inversion techniques, often leading to structural degradation during editing. In this paper, we introduce GIDE (Grounded Inversion for DLLM Image Editing), a unified framework designed to bridge this gap. GIDE incorporates a novel Discrete Noise Inversion mechanism that accurately captures latent noise patterns within the discrete token space, ensuring high-fidelity reconstruction. We then decompose the editing pipeline into grounding, inversion, and refinement stages. This design enables GIDE supporting various editing instructions (text, point and box) and operations while strictly preserving the unedited background. Furthermore, to overcome the limitations of existing single-step evaluation protocols, we introduce GIDE-Bench, a rigorous benchmark comprising 805 compositional editing scenarios guided by diverse multi-modal inputs. Extensive experiments on GIDE-Bench demonstrate that GIDE significantly outperforms prior training-free methods, improving Semantic Correctness by 51.83% and Perceptual Quality by 50.39%. Additional evaluations on ImgEdit-Bench confirm its broad applicability, demonstrating consistent gains over trained baselines and yielding photorealistic consistency on par with leading models.

AI agents capable of controlling user interfaces have the potential to transform human interaction with digital devices. To accelerate this transformation, two fundamental building blocks are essential: high-quality datasets that enable agents to achieve complex and human-relevant goals, and robust evaluation methods that allow researchers and practitioners to rapidly enhance agent performance. In this paper, we introduce DigiData, a large-scale, high-quality, diverse, multi-modal dataset designed for training mobile control agents. Unlike existing datasets, which derive goals from unstructured interactions, DigiData is meticulously constructed through comprehensive exploration of app features, resulting in greater diversity and higher goal complexity. Additionally, we present DigiData-Bench, a benchmark for evaluating mobile control agents on real-world complex tasks. We demonstrate that the commonly used step-accuracy metric falls short in reliably assessing mobile control agents and, to address this, we propose dynamic evaluation protocols and AI-powered evaluations as rigorous alternatives for agent assessment. Our contributions aim to significantly advance the development of mobile control agents, paving the way for more intuitive and effective human-device interactions.

Post-wildfire debris flows (PFDFs) are destructive sediment-laden hazards triggered when intense rainfall strikes recently burned terrain, destabilizing hillslopes and threatening infrastructure, local economies, and community safety. Data-driven methods have been proposed to learn predictive patterns directly from historical PFDF observations. However, the current research landscape of data-driven PFDF prediction remains highly fragmented across feature spaces, model architectures, and evaluation protocols, making rigorous comparison and the derivation of scientific insights difficult. Moreover, existing studies lack a systematic investigation into the relative importance of heterogeneous factors (e.g., meteorological conditions, terrain characteristics, soil properties, and burn severity) in triggering PFDF. To address these limitations, we present a unified benchmark for data-driven PFDF prediction, enabling fair and comprehensive evaluation across diverse models and feature configurations. Furthermore, to better understand the underlying drivers of PFDF formation, we propose a reinforcement learning-based feature selection framework that identifies factors whose perturbations render positive and negative events indistinguishable, thereby discovering the regional underlying mechanisms of PFDF occurrence across regions. Our code and benchmark are publicly available at https://github.com/KINDLab-Fly/PFDF-Benchmark.

Large Language Models (LLMs) are increasingly integrated into real-world applications via the Model Context Protocol (MCP), a universal, open standard for connecting AI agents with data sources and external tools. While MCP enhances the capabilities of LLM-based agents, it also introduces new security risks and expands their attack surfaces. In this paper, we present the first systematic taxonomy of MCP security, identifying 17 attack types across 4 primary attack surfaces. We introduce MCPSecBench, a comprehensive security benchmark and playground that integrates prompt datasets, MCP servers, MCP clients, attack scripts, and protection mechanisms to evaluate these attacks across three major MCP providers. Our benchmark is modular and extensible, allowing researchers to incorporate custom implementations of clients, servers, and transport protocols for systematic security assessment. Experimental results show that over 85% of the identified attacks successfully compromise at least one platform, with core vulnerabilities universally affecting Claude, OpenAI, and Cursor, while prompt-based and tool-centric attacks exhibit considerable variability across different hosts and models. In addition, current protection mechanisms have little effect against these attacks. Overall, MCPSecBench standardizes the evaluation of MCP security and enables rigorous testing across all MCP layers.

In this paper, we present a comprehensive overview of the NTIRE 2026 3rd Restore Any Image Model (RAIM) challenge, with a specific focus on Track 3: AI Flash Portrait. Despite significant advancements in deep learning for image restoration, existing models still encounter substantial challenges in real-world low-light portrait scenarios. Specifically, they struggle to achieve an optimal balance among noise suppression, detail preservation, and faithful illumination and color reproduction. To bridge this gap, this challenge aims to establish a novel benchmark for real-world low-light portrait restoration. We comprehensively evaluate the proposed algorithms utilizing a hybrid evaluation system that integrates objective quantitative metrics with rigorous subjective assessment protocols. For this competition, we provide a dataset containing 800 groups of real-captured low-light portrait data. Each group consists of a 1K-resolution low-light input image, a 1K ground truth (GT), and a 1K person mask. This challenge has garnered widespread attention from both academia and industry, attracting over 100 participating teams and receiving more than 3,000 valid submissions. This report details the motivation behind the challenge, the dataset construction process, the evaluation metrics, and the various phases of the competition. The released dataset and baseline code for this track are publicly available from the same https://github.com/zsn1434/AI_Flash-BaseLine/tree/main{GitHub repository}, and the official challenge webpage is hosted on https://www.codabench.org/competitions/12885/{CodaBench}.

Large Language Models are routinely compressed via post-training quantization to reduce inference costs and memory footprint for cloud and edge deployment, yet the impact of this compression on model quality remains poorly understood. Existing studies typically compare only two conditions (full-precision vs. a single quantized variant), rely on aggregate bias metrics, and evaluate a single model family, making it impossible to distinguish gradual degradation from threshold-dependent safety failures. We conduct a controlled empirical study of three instruction-tuned models (Qwen2.5-7B, Mistral-7B, Phi-3.5-mini) at five precision levels (BF16 through 3-bit) on 12,148 BBQ bias benchmark items across 5 random seeds, totaling 911,100 inference records. Our results reveal that 3-bit quantization causes 6-21% of previously unbiased items to develop new stereotypical behaviors, following a clear dose-response pattern confirmed via logistic regression, while models' willingness to select "unknown" answers declines by 17.4%. Crucially, these item-level changes are invisible to standard quality metrics: perplexity increases by less than 0.5% at 8-bit and under 3% at 4-bit across all three models, yet 2.5-5.6% of items already develop new biases at 4-bit. These findings demonstrate that aggregate evaluation metrics systematically miss fairness-critical degradation, underscoring the need for quality-aware compression protocols that explicitly test for bias emergence before deployment.

Deep search enables language agents to solve complex information-seeking tasks through iterative web exploration. However, developing robust open-source agents remains challenging for three reasons. First, opaque trajectory-construction pipelines make training data difficult to audit and reproduce. Second, context compression can create a substantial mismatch between training and inference. Third, agents may retrieve benchmark content from the open web during evaluation, contaminating the results and overstating their performance. To overcome these bottlenecks, we introduce Glasseek, a fully transparent deep search framework spanning the complete development lifecycle, releasing all data, model weights, training recipes, and code. Specifically, for data construction, we establish a verifiable trajectory curation pipeline that systematically generates high-quality exploration paths. For training, we propose Compaction-Managed GSPO (CM-GSPO), which uses the same policy for search and summarization while optimizing only execution sub-trajectories, enabling end-to-end search-policy training across context-compaction boundaries. For evaluation, we conduct the first systematic audit of test-set leakage in open-web agents and release a modular anti-leakage plugin that seamlessly integrates into existing evaluation harnesses. Under this rigorous evaluation protocol, our 9B agent outperforms frontier proprietary agents across almost all deep research benchmarks spanning diverse task types, and achieves the best overall performance among recent open-weight agents, setting a new state of the art at the 9B scale.

The performance of speaker verification systems degrades significantly under language mismatch, a critical challenge exacerbated by the field's reliance on English-centric data. To address this, we propose the TidyVoice Challenge for cross-lingual speaker verification. The challenge leverages the TidyVoiceX dataset from the novel TidyVoice benchmark, a large-scale, multilingual corpus derived from Mozilla Common Voice, and specifically curated to isolate the effect of language switching across approximately 40 languages. Participants will be tasked with building systems robust to this mismatch, with performance primarily evaluated using the Equal Error Rate on cross-language trials. By providing standardized data, open-source baselines, and a rigorous evaluation protocol, this challenge aims to drive research towards fairer, more inclusive, and language-independent speaker recognition technologies, directly aligning with the Interspeech 2026 theme, "Speaking Together."

Speech quality predictors are increasingly used as rewards, yet no agreed measure of their hackability exists. The usual measurement has two flaws. First, the perturbation reaches the predictor through a processing chain -- here a neural codec -- that shifts the score on its own, which scoring against the raw input charges to the attack. Referencing the unperturbed round trip instead changes measured hackability by up to a factor of four (0.31 to 0.08 for one defence). Second, one trained attacker is a sample, not a measurement: five attackers differing only in random seed reach success rates from 0.00 to 0.38 against one fixed predictor, so a defence claim needs the worst case over several. Under this protocol, four published predictors differ widely: NISQA is hacked on 90% of utterances, SSL-MOS on 21%, DNSMOS on 14% and UTMOS on 6%. We then audit a closed attack-detect-patch loop. It hardens the predictor only in its own attack space, by less than the spread between attackers; a random-perturbation baseline matches it; and it costs up to 0.30 system SRCC out of domain. Enhancers post-trained against patched predictors hack them far less (PESQ -0.03 versus -0.23). Code, preregistration and run outputs are released.

Deep research systems represent an emerging class of agentic information retrieval methods that generate comprehensive and well-supported reports to complex queries. However, most existing frameworks rely on dynamic commercial search APIs, which pose reproducibility and transparency challenges in addition to their cost. To address these limitations, we introduce DeepResearchGym, an open-source sandbox that combines a reproducible search API with a rigorous evaluation protocol for benchmarking deep research systems. The API indexes large-scale public web corpora, namely ClueWeb22 and FineWeb, using a state-of-the-art dense retriever and approximate nearest neighbor search via DiskANN. It achieves lower latency than popular commercial APIs while ensuring stable document rankings across runs, and is freely available for research use. To evaluate deep research systems' outputs, we extend the Researchy Questions benchmark with automatic metrics through LLM-as-a-judge assessments to measure alignment with users' information needs, retrieval faithfulness, and report quality. Experimental results show that systems integrated with DeepResearchGym achieve performance comparable to those using commercial APIs, with performance rankings remaining consistent across evaluation metrics. A human evaluation study further confirms that our automatic protocol aligns with human preferences, validating the framework's ability to help support controlled assessment of deep research systems. Our code and API documentation are available at https://www.deepresearchgym.ai.

Recent advancements in image manipulation have achieved unprecedented progress in generating photorealistic content, but also simultaneously eliminating barriers to arbitrary manipulation and editing, raising concerns about multimedia authenticity and cybersecurity. However, existing Image Manipulation Detection and Localization (IMDL) methodologies predominantly focus on splicing or copy-move forgeries, lacking dedicated benchmarks for inpainting-based manipulations. To bridge this gap, we present COCOInpaint, a comprehensive benchmark specifically designed for inpainting detection, with three key contributions: 1) High-quality inpainting samples generated by six state-of-the-art inpainting models, 2) Diverse generation scenarios enabled by four mask generation strategies with optional text guidance, and 3) Large-scale coverage with 258,266 inpainted images with rich semantic diversity. Our benchmark is constructed to emphasize intrinsic inconsistencies between inpainted and authentic regions, rather than superficial semantic artifacts such as object shapes. We establish a rigorous evaluation protocol using three standard metrics to assess existing IMDL approaches. The dataset will be made publicly available to facilitate future research in this area.

Unified multimodal models target joint understanding, reasoning, and generation, but current image editing benchmarks are largely confined to natural images and shallow commonsense reasoning, offering limited assessment of this capability under structured, domain-specific constraints. In this work, we introduce GRADE, the first benchmark to assess discipline-informed knowledge and reasoning in image editing. GRADE comprises 520 carefully curated samples across 10 academic domains, spanning from natural science to social science. To support rigorous evaluation, we propose a multi-dimensional evaluation protocol that jointly assesses Discipline Reasoning, Visual Consistency, and Logical Readability. Extensive experiments on 20 state-of-the-art open-source and closed-source models reveal substantial limitations in current models under implicit, knowledge-intensive editing settings, leading to large performance gaps. Beyond quantitative scores, we conduct rigorous analyses and ablations to expose model shortcomings and identify the constraints within disciplinary editing. Together, GRADE pinpoints key directions for the future development of unified multimodal models, advancing the research on discipline-informed image editing and reasoning. Our benchmark and evaluation code are publicly released.

In this paper, we introduce CineBrain, the first large-scale dataset featuring simultaneous EEG and fMRI recordings during dynamic audiovisual stimulation. Recognizing the complementary strengths of EEG's high temporal resolution and fMRI's deep-brain spatial coverage, CineBrain provides approximately six hours of narrative-driven content from the popular television series The Big Bang Theory for each of six participants. Building upon this unique dataset, we propose CineSync, an innovative multimodal decoding framework integrates a Multi-Modal Fusion Encoder with a diffusion-based Neural Latent Decoder. Our approach effectively fuses EEG and fMRI signals, significantly improving the reconstruction quality of complex audiovisual stimuli. To facilitate rigorous evaluation, we introduce Cine-Benchmark, a comprehensive evaluation protocol that assesses reconstructions across semantic and perceptual dimensions. Experimental results demonstrate that CineSync achieves state-of-the-art video reconstruction performance and highlight our initial success in combining fMRI and EEG for reconstructing both video and audio stimuli. Project Page: https://jianxgao.github.io/CineBrain.

We present InterHandGen, a novel framework that learns the generative prior of two-hand interaction. Sampling from our model yields plausible and diverse two-hand shapes in close interaction with or without an object. Our prior can be incorporated into any optimization or learning methods to reduce ambiguity in an ill-posed setup. Our key observation is that directly modeling the joint distribution of multiple instances imposes high learning complexity due to its combinatorial nature. Thus, we propose to decompose the modeling of joint distribution into the modeling of factored unconditional and conditional single instance distribution. In particular, we introduce a diffusion model that learns the single-hand distribution unconditional and conditional to another hand via conditioning dropout. For sampling, we combine anti-penetration and classifier-free guidance to enable plausible generation. Furthermore, we establish the rigorous evaluation protocol of two-hand synthesis, where our method significantly outperforms baseline generative models in terms of plausibility and diversity. We also demonstrate that our diffusion prior can boost the performance of two-hand reconstruction from monocular in-the-wild images, achieving new state-of-the-art accuracy.

Building a world model requires a state representation useful for planning and decision-making---potentially over tasks unknown at training time. In the context of financial markets, planning and decision-making may require a model to reason about market-wide conditions, asset-specific expected returns, liquidity, volatility, and cross-asset relationships. Yet financial representation learning has largely been evaluated on individual predictive tasks, oftentimes on a single time period using comparatively narrow datasets. We address this through three primary contributions. First, we introduce Market-1T, a dataset containing nearly one trillion observations across U.S. equities from 2008 to 2025 at 1 Hz resolution. Second, we develop and implement a rigorous evaluation protocol. Third, we conduct a systematic large-scale study of financial representation learning, comparing 18 encoder-training strategies across nearly two decades of market regimes. We evaluate learned representations both by their predictive utility on common finance tasks and through probes of latent structure. We find that encoders with similar predictive performance can organize market state very differently. Collectively, we establish a foundation for training and evaluating financial market representations in support of world models such as DINO-WM, V-JEPA 2, and LeWM.

Large language models (LLMs) have demonstrated remarkable capabilities, but they still frequently produce hallucinations. These hallucinations are difficult to detect in reasoning-intensive tasks, where the content appears coherent but contains errors like logical flaws and unreliable intermediate results. While step-level analysis is commonly used to detect internal hallucinations, it suffers from limited granularity and poor scalability due to its reliance on step segmentation. To address these limitations, we propose TokenHD, a holistic pipeline for training token-level hallucination detectors. Specifically, TokenHD consists of a scalable data engine for synthesizing large-scale hallucination annotations along with a training recipe featuring an importance-weighted strategy for robust model training. To systematically assess the detection performance, we also provide a rigorous evaluation protocol. Through training within TokenHD, our detector operates directly on free-form text to identify hallucinations, eliminating the need for predefined step segmentation or additional text reformatting. Our experiments show that even a small detector (0.6B) achieves substantial performance gains after training, surpassing much larger reasoning models (e.g., QwQ-32B), and detection performance scales consistently with model size from 0.6B to 8B. Finally, we show that our detector can generalize well across diverse practical scenarios and explore strategies to further enhance its cross-domain generalization capability.

Data scaling has revolutionized fields like natural language processing and computer vision, providing models with remarkable generalization capabilities. In this paper, we investigate whether similar data scaling laws exist in robotics, particularly in robotic manipulation, and whether appropriate data scaling can yield single-task robot policies that can be deployed zero-shot for any object within the same category in any environment. To this end, we conduct a comprehensive empirical study on data scaling in imitation learning. By collecting data across numerous environments and objects, we study how a policy's generalization performance changes with the number of training environments, objects, and demonstrations. Throughout our research, we collect over 40,000 demonstrations and execute more than 15,000 real-world robot rollouts under a rigorous evaluation protocol. Our findings reveal several intriguing results: the generalization performance of the policy follows a roughly power-law relationship with the number of environments and objects. The diversity of environments and objects is far more important than the absolute number of demonstrations; once the number of demonstrations per environment or object reaches a certain threshold, additional demonstrations have minimal effect. Based on these insights, we propose an efficient data collection strategy. With four data collectors working for one afternoon, we collect sufficient data to enable the policies for two tasks to achieve approximately 90% success rates in novel environments with unseen objects.

Vision-Language Models (VLMs) have demonstrated remarkable progress in multimodal understanding, yet their capabilities for scientific reasoning remains inadequately assessed. Current multimodal benchmarks predominantly evaluate generic image comprehension or text-driven reasoning, lacking authentic scientific contexts that require domain-specific knowledge integration with visual evidence analysis. To fill this gap, we present CSVQA, a diagnostic multimodal benchmark specifically designed for evaluating scientific reasoning through domain-grounded visual question answering.Our benchmark features 1,378 carefully constructed question-answer pairs spanning diverse STEM disciplines, each demanding domain knowledge, integration of visual evidence, and higher-order reasoning. Compared to prior multimodal benchmarks, CSVQA places greater emphasis on real-world scientific content and complex reasoning.We additionally propose a rigorous evaluation protocol to systematically assess whether model predictions are substantiated by valid intermediate reasoning steps based on curated explanations. Our comprehensive evaluation of 15 VLMs on this benchmark reveals notable performance disparities, as even the top-ranked proprietary model attains only 49.6\% accuracy.This empirical evidence underscores the pressing need for advancing scientific reasoning capabilities in VLMs. Our CSVQA is released at https://huggingface.co/datasets/Skywork/CSVQA.

In medical imaging, vision-language models face a critical duality: pretrained networks offer broad robustness but lack subtle, modality-specific characteristics, while fine-tuned expert models achieve high in-distribution accuracy yet falter under modality shift. Existing model-merging techniques, designed for natural-image benchmarks, are simple and efficient but fail to deliver consistent gains across diverse medical modalities; their static interpolation limits reliability in varied clinical tasks. To address this, we introduce Test-Time Task adaptive merging (T^3), a backpropagation-free framework that computes per-sample interpolation coefficients via the Jensen-Shannon divergence between the two models' output distributions. T^3 dynamically preserves local precision when models agree and defers to generalist robustness under drift. To overcome the inference costs of sample-wise merging, we further propose a batch-wise extension, T^3_B, that computes a merging coefficient across a batch of samples, dramatically reducing computational bottleneck. Recognizing the lack of a standardized medical-merging benchmark, we present a rigorous cross-evaluation protocol spanning in-domain, base-to-novel, and corruptions across four modalities. Empirically, T^3 sets new state-of-the-art in Top-1 accuracy and error reduction, outperforming strong baselines while maintaining efficiency, paving the way for adaptive MVLM deployment in clinical settings. Our code is available at https://github.com/Razaimam45/TCube.

We introduce ESGenius, a comprehensive benchmark for evaluating and enhancing the proficiency of Large Language Models (LLMs) in Environmental, Social and Governance (ESG) and sustainability-focused question answering. ESGenius comprises two key components: (i) ESGenius-QA, a collection of 1 136 multiple-choice questions generated by LLMs and rigorously validated by domain experts, covering a broad range of ESG pillars and sustainability topics. Each question is systematically linked to its corresponding source text, enabling transparent evaluation and supporting retrieval-augmented generation (RAG) methods; and (ii) ESGenius-Corpus, a meticulously curated repository of 231 foundational frameworks, standards, reports and recommendation documents from seven authoritative sources. Moreover, to fully assess the capabilities and adaptation potential of the model, we implement a rigorous two-stage evaluation protocol -- Zero-Shot and RAG. Extensive experiments across 50 LLMs (ranging from 0.5 B to 671 B parameters) demonstrate that state-of-the-art models achieve only moderate performance in zero-shot settings, with accuracies typically around 55--70\%, highlighting ESGenius's challenging nature for LLMs in interdisciplinary contexts. However, models employing RAG show significant performance improvements, particularly for smaller models. For example, "DeepSeek-R1-Distill-Qwen-14B" improves from 63.82\% (zero-shot) to 80.46\% with RAG. These results underscore the necessity of grounding responses in authoritative sources for enhanced ESG understanding. To the best of our knowledge, ESGenius is the first benchmark curated for LLMs and the relevant enhancement technologies that focuses on ESG and sustainability topics.

Understanding and reasoning over text within visual contexts poses a significant challenge for Vision-Language Models (VLMs), given the complexity and diversity of real-world scenarios. To address this challenge, text-rich Visual Question Answering (VQA) datasets and benchmarks have emerged for high-resource languages like English. However, a critical gap persists for low-resource languages such as Korean, where the lack of comprehensive benchmarks hinders robust model evaluation and comparison. To bridge this gap, we introduce KRETA, a benchmark for Korean Reading and rEasoning in Text-rich VQA Attuned to diverse visual contexts. KRETA facilitates an in-depth evaluation of both visual text understanding and reasoning capabilities, while also supporting a multifaceted assessment across 15 domains and 26 image types. Additionally, we introduce a semi-automated VQA generation pipeline specifically optimized for text-rich settings, leveraging refined stepwise image decomposition and a rigorous seven-metric evaluation protocol to ensure data quality. While KRETA is tailored for Korean, we hope our adaptable and extensible pipeline will facilitate the development of similar benchmarks in other languages, thereby accelerating multilingual VLM research. The code and dataset for KRETA are available at https://github.com/tabtoyou/KRETA.

Deepfake detection remains a formidable challenge due to the complex and evolving nature of fake content in real-world scenarios. However, existing academic benchmarks suffer from severe discrepancies from industrial practice, typically featuring homogeneous training sources and low-quality testing images, which hinder the practical deployments of current detectors. To mitigate this gap, we introduce HydraFake, a dataset that simulates real-world challenges with hierarchical generalization testing. Specifically, HydraFake involves diversified deepfake techniques and in-the-wild forgeries, along with rigorous training and evaluation protocol, covering unseen model architectures, emerging forgery techniques and novel data domains. Building on this resource, we propose Veritas, a multi-modal large language model (MLLM) based deepfake detector. Different from vanilla chain-of-thought (CoT), we introduce pattern-aware reasoning that involves critical reasoning patterns such as "planning" and "self-reflection" to emulate human forensic process. We further propose a two-stage training pipeline to seamlessly internalize such deepfake reasoning capacities into current MLLMs. Experiments on HydraFake dataset reveal that although previous detectors show great generalization on cross-model scenarios, they fall short on unseen forgeries and data domains. Our Veritas achieves significant gains across different OOD scenarios, and is capable of delivering transparent and faithful detection outputs.

Transforming static research papers into dynamic media such as posters, slides, and videos is essential for effective dissemination but remains a labor-intensive challenge. Existing automated approaches often treat these formats in isolation and consequently fail to maintain semantic consistency across the entire presentation suite. We address this fragmentation by formalizing the task of unified presentation suite generation and proposing OmniPresent to orchestrate the creation of coherent deliverables. Our framework adopts a renderable HTML representation to enable centralized content planning and a self-correcting verify-and-repair loop that actively resolves conflicts across modalities. We further facilitate scalable research in this domain by releasing OmniPreBench, a comprehensive dataset comprising over one thousand papers with paired artifacts, and establishing a rigorous VLM-based evaluation protocol. Empirical results confirm that our method generates high-quality and faithful presentation suites that significantly surpass strong baselines in both accuracy and visual appeal.

Combinatorics is central to Olympiad-level mathematical problem solving, requiring deep discrete reasoning, creative constructions, and rigorous structural insight. Recent evidence suggests that even today's strongest frontier models remain uneven on Olympiad combinatorics, revealing a gap in creative mathematical reasoning. We introduce ComBench, an Olympiad-level combinatorics benchmark for evaluating and diagnosing the combinatorial reasoning capabilities of large language models. ComBench contains 100 human-annotated competition-level problems organized around two complementary settings: analysis-centric problems, which primarily require rigorous mathematical arguments, and construction-centric problems, which require explicit constructions in addition to correctness justifications. The evaluation protocol combines rubric-guided proof grading with deterministic construction verification, exposing cases where proof quality and construction validity diverge. Experiments on frontier open- and closed-source models show that ComBench is far from saturated: the strongest model reaches 65.4% overall Avg. and 75.3% overall Best@4. We further find that Rigorous Proof Reasoning and Constructive Realization are distinct capabilities: Kimi-K2.6 trails GPT-5.5 on analysis-centric proof grading but surpasses it on construction-centric Best@4, while Existence and Construction problems remain consistently hardest across representative frontier models.

The Model Context Protocol has emerged as a transformative standard for connecting large language models to external data sources and tools, rapidly gaining adoption across major AI providers and development platforms. However, existing benchmarks are overly simplistic and fail to capture real application challenges such as long-horizon reasoning and large, unfamiliar tool spaces. To address this critical gap, we introduce MCP-Universe, the first comprehensive benchmark specifically designed to evaluate LLMs in realistic and hard tasks through interaction with real-world MCP servers. Our benchmark encompasses 6 core domains spanning 11 different MCP servers: Location Navigation, Repository Management, Financial Analysis, 3D Design, Browser Automation, and Web Searching. To ensure rigorous evaluation, we implement execution-based evaluators, including format evaluators for agent format compliance, static evaluators for time-invariant content matching, and dynamic evaluators that automatically retrieve real-time ground truth for temporally sensitive tasks. Through extensive evaluation of leading LLMs, we find that even SOTA models such as GPT-5 (43.72%), Grok-4 (33.33%) and Claude-4.0-Sonnet (29.44%) exhibit significant performance limitations. In addition, our benchmark poses a significant long-context challenge for LLM agents, as the number of input tokens increases rapidly with the number of interaction steps. Moreover, it introduces an unknown-tools challenge, as LLM agents often lack familiarity with the precise usage of the MCP servers. Notably, enterprise-level agents like Cursor cannot achieve better performance than standard ReAct frameworks. Beyond evaluation, we open-source our extensible evaluation framework with UI support, enabling researchers and practitioners to seamlessly integrate new agents and MCP servers while fostering innovation in the rapidly evolving MCP ecosystem.

The integration of Large Language Models (LLMs) with Internet-of-Things (IoT) systems faces significant challenges in hardware heterogeneity and control complexity. The Model Context Protocol (MCP) emerges as a critical enabler, providing standardized communication between LLMs and physical devices. We propose IoT-MCP, a novel framework that implements MCP through edge-deployed servers to bridge LLMs and IoT ecosystems. To support rigorous evaluation, we introduce IoT-MCP Bench, the first benchmark containing 114 Basic Tasks (e.g., ``What is the current temperature?'') and 1,140 Complex Tasks (e.g., ``I feel so hot, do you have any ideas?'') for IoT-enabled LLMs. Experimental validation across 22 sensor types and 6 microcontroller units demonstrates IoT-MCP's 100% task success rate to generate tool calls that fully meet expectations and obtain completely accurate results, 205ms average response time, and 74KB peak memory footprint. This work delivers both an open-source integration framework (https://github.com/Duke-CEI-Center/IoT-MCP-Servers) and a standardized evaluation methodology for LLM-IoT systems.

The deployment of Large Language Models (LLMs) for structuring clinical data is critically hindered by their tendency to hallucinate facts and their inability to follow domain-specific rules. To address this, we introduce MedPAO, a novel agentic framework that ensures accuracy and verifiable reasoning by grounding its operation in established clinical protocols such as the ABCDEF protocol for CXR analysis. MedPAO decomposes the report structuring task into a transparent process managed by a Plan-Act-Observe (PAO) loop and specialized tools. This protocol-driven method provides a verifiable alternative to opaque, monolithic models. The efficacy of our approach is demonstrated through rigorous evaluation: MedPAO achieves an F1-score of 0.96 on the critical sub-task of concept categorization. Notably, expert radiologists and clinicians rated the final structured outputs with an average score of 4.52 out of 5, indicating a level of reliability that surpasses baseline approaches relying solely on LLM-based foundation models. The code is available at: https://github.com/MiRL-IITM/medpao-agent

The integration of Large Language Models (LLMs) into software engineering has shifted the focus from function-level generation to repository-scale assistance. However, existing benchmarks largely rely on bug reports from GitHub Issues, which often allow models to bypass genuine understanding via pattern matching on error logs. This misalignment under-measures Edit Bias, which refers to premature generation, where models prematurely propose code modifications instead of understanding the existing repository architecture. Furthermore, current LLM-as-a-Judge scalar scoring suffers from high variance and low interpretability. This work introduces RepoProbe, a novel benchmark for evaluating repository-level code understanding through open-ended Q&A using GitHub Discussions, which focuses on open-ended architectural inquiries rather than defect reporting. To ensure rigorous evaluation, we propose a Checklist-Based Verification Protocol that decomposes answers into atomic, verifiable facts, thereby replacing subjective ratings with objective verification. Our evaluation of state-of-the-art (SOTA) LLMs reveals a persistent gap between high clarity and evidencegrounded technical correctness. It also quantitatively confirms the prevalence of edit bias, in which models prioritize code generation instead of architectural analysis. Finally, we demonstrate that our verification protocol significantly improves evaluation reliability compared to traditional evaluations with scalar scoring.

We present Yambda-5B, a large-scale open dataset sourced from the Yandex.Music streaming platform. Yambda-5B contains 4.79 billion user-item interactions from 1 million users across 9.39 million tracks. The dataset includes two primary types of interactions: implicit feedback (listening events) and explicit feedback (likes, dislikes, unlikes and undislikes). In addition, we provide audio embeddings for most tracks, generated by a convolutional neural network trained on audio spectrograms. A key distinguishing feature of Yambda-5B is the inclusion of the is_organic flag, which separates organic user actions from recommendation-driven events. This distinction is critical for developing and evaluating machine learning algorithms, as Yandex.Music relies on recommender systems to personalize track selection for users. To support rigorous benchmarking, we introduce an evaluation protocol based on a Global Temporal Split, allowing recommendation algorithms to be assessed in conditions that closely mirror real-world use. We report benchmark results for standard baselines (ItemKNN, iALS) and advanced models (SANSA, SASRec) using a variety of evaluation metrics. By releasing Yambda-5B to the community, we aim to provide a readily accessible, industrial-scale resource to advance research, foster innovation, and promote reproducible results in recommender systems.

Following instructions in real-world conditions requires the ability to adapt to the world's volatility and entanglement: the environment is dynamic and unpredictable, instructions can be linguistically complex with diverse vocabulary, and the number of possible goals an agent may encounter is vast. Despite extensive research in this area, most studies are conducted in static environments with simple instructions and a limited vocabulary, making it difficult to assess agent performance in more diverse and challenging settings. To address this gap, we introduce CrafText, a benchmark for evaluating instruction following in a multimodal environment with diverse instructions and dynamic interactions. CrafText includes 3,924 instructions with 3,423 unique words, covering Localization, Conditional, Building, and Achievement tasks. Additionally, we propose an evaluation protocol that measures an agent's ability to generalize to novel instruction formulations and dynamically evolving task configurations, providing a rigorous test of both linguistic understanding and adaptive decision-making.

We tackle open-world semantic segmentation, which aims at learning to segment arbitrary visual concepts in images, by using only image-text pairs without dense annotations. Existing open-world segmentation methods have shown impressive advances by employing contrastive learning (CL) to learn diverse visual concepts and transferring the learned image-level understanding to the segmentation task. However, these CL-based methods suffer from a train-test discrepancy, since it only considers image-text alignment during training, whereas segmentation requires region-text alignment during testing. In this paper, we proposed a novel Text-grounded Contrastive Learning (TCL) framework that enables a model to directly learn region-text alignment. Our method generates a segmentation mask for a given text, extracts text-grounded image embedding from the masked region, and aligns it with text embedding via TCL. By learning region-text alignment directly, our framework encourages a model to directly improve the quality of generated segmentation masks. In addition, for a rigorous and fair comparison, we present a unified evaluation protocol with widely used 8 semantic segmentation datasets. TCL achieves state-of-the-art zero-shot segmentation performances with large margins in all datasets. Code is available at https://github.com/kakaobrain/tcl.

The real world is dynamic, yet most image fusion methods process static frames independently, ignoring temporal correlations in videos and leading to flickering and temporal inconsistency. To address this, we propose Unified Video Fusion (UniVF), a novel framework for temporally coherent video fusion that leverages multi-frame learning and optical flow-based feature warping for informative, temporally coherent video fusion. To support its development, we also introduce Video Fusion Benchmark (VF-Bench), the first comprehensive benchmark covering four video fusion tasks: multi-exposure, multi-focus, infrared-visible, and medical fusion. VF-Bench provides high-quality, well-aligned video pairs obtained through synthetic data generation and rigorous curation from existing datasets, with a unified evaluation protocol that jointly assesses the spatial quality and temporal consistency of video fusion. Extensive experiments show that UniVF achieves state-of-the-art results across all tasks on VF-Bench. Project page: https://vfbench.github.io.

While Large Language Models (LLMs) have achieved remarkable performance, they remain vulnerable to jailbreak. The integration of Large Language Models (LLMs) with external tools via protocols such as the Model Context Protocol (MCP) introduces critical security vulnerabilities, including prompt injection, data exfiltration, and other threats. To counter these challenges, we propose MCP-GUARD, a robust, layered defense architecture designed for LLM-tool interactions. MCP-GUARD employs a three-stage detection pipeline that balances efficiency with accuracy: it progresses from lightweight static scanning for overt threats and a deep neural detector for semantic attacks, to our fine-tuned E5-based model which achieves 96.01\% accuracy in identifying adversarial prompts. Finally, an LLM arbitrator synthesizes these signals to deliver the final decision. To enable rigorous training and evaluation, we introduce MCP-ATTACKBENCH, a comprehensive benchmark comprising 70,448 samples augmented by GPT-4. This benchmark simulates diverse real-world attack vectors that circumvent conventional defenses in the MCP paradigm, thereby laying a solid foundation for future research on securing LLM-tool ecosystems.

While Large Language Models (LLMs) have demonstrated high proficiency on English-centric medical examinations, their performance often declines when faced with non-English languages and multimodal diagnostic tasks. This study protocol describes the development of EuropeMedQA, the first comprehensive, multilingual, and multimodal medical examination dataset sourced from official regulatory exams in Italy, France, Spain, and Portugal. Following FAIR data principles and SPIRIT-AI guidelines, we describe a rigorous curation process and an automated translation pipeline for comparative analysis. We evaluate contemporary multimodal LLMs using a zero-shot, strictly constrained prompting strategy to assess cross-lingual transfer and visual reasoning. EuropeMedQA aims to provide a contamination-resistant benchmark that reflects the complexity of European clinical practices and fosters the development of more generalizable medical AI.

Molecular function is largely determined by structure. Accurately aligning molecular structure with natural language is therefore essential for enabling large language models (LLMs) to reason about downstream chemical tasks. However, the substantial cost of human annotation makes it infeasible to construct large-scale, high-quality datasets of structure-grounded descriptions. In this work, we propose a fully automated annotation framework for generating precise molecular structure descriptions at scale. Our approach builds upon and extends a rule-based chemical nomenclature parser to interpret IUPAC names and construct enriched, structured XML metadata that explicitly encodes molecular structure. This metadata is then used to guide LLMs in producing accurate natural-language descriptions. Using this framework, we curate a large-scale dataset of approximately 163k molecule-description pairs. A rigorous validation protocol combining LLM-based and expert human evaluation on a subset of 2,000 molecules demonstrates a high description precision of 98.6%. The resulting dataset provides a reliable foundation for future molecule-language alignment, and the proposed annotation method is readily extensible to larger datasets and broader chemical tasks that rely on structural descriptions.

Current Vision Language Models (VLMs) demonstrate a critical gap between surface-level recognition and deep narrative reasoning when processing sequential visual storytelling. Through a comprehensive investigation of manga narrative understanding, we reveal that while recent large multimodal models excel at individual panel interpretation, they systematically fail at temporal causality and cross-panel cohesion, core requirements for coherent story comprehension. We introduce a novel evaluation framework that combines fine-grained multimodal annotation, cross-modal embedding analysis, and retrieval-augmented assessment to systematically characterize these limitations. Our methodology includes (i) a rigorous annotation protocol linking visual elements to narrative structure through aligned light novel text, (ii) comprehensive evaluation across multiple reasoning paradigms, including direct inference and retrieval-augmented generation, and (iii) cross-modal similarity analysis revealing fundamental misalignments in current VLMs' joint representations. Applying this framework to Re:Zero manga across 11 chapters with 308 annotated panels, we conduct the first systematic study of long-form narrative understanding in VLMs through three core evaluation axes: generative storytelling, contextual dialogue grounding, and temporal reasoning. Our findings demonstrate that current models lack genuine story-level intelligence, struggling particularly with non-linear narratives, character consistency, and causal inference across extended sequences. This work establishes both the foundation and practical methodology for evaluating narrative intelligence, while providing actionable insights into the capability of deep sequential understanding of Discrete Visual Narratives beyond basic recognition in Multimodal Models. Project Page: https://re-verse.vercel.app

Recent text-to-video (T2V) technology advancements, as demonstrated by models such as Gen2, Pika, and Sora, have significantly broadened its applicability and popularity. Despite these strides, evaluating these models poses substantial challenges. Primarily, due to the limitations inherent in automatic metrics, manual evaluation is often considered a superior method for assessing T2V generation. However, existing manual evaluation protocols face reproducibility, reliability, and practicality issues. To address these challenges, this paper introduces the Text-to-Video Human Evaluation (T2VHE) protocol, a comprehensive and standardized protocol for T2V models. The T2VHE protocol includes well-defined metrics, thorough annotator training, and an effective dynamic evaluation module. Experimental results demonstrate that this protocol not only ensures high-quality annotations but can also reduce evaluation costs by nearly 50%. We will open-source the entire setup of the T2VHE protocol, including the complete protocol workflow, the dynamic evaluation component details, and the annotation interface code. This will help communities establish more sophisticated human assessment protocols.

AI pentesting agents are increasingly credible as offensive security systems, but current benchmarks still provide limited guidance on which will perform best in real-world targets. Existing evaluation protocols assess and optimize for predefined goals such as capture-the-flag, remote code execution, exploit reproduction, or trajectory similarity, in simplified or narrow settings. These tools are valuable for measuring bounded capabilities, yet they do not adequately capture the complexity, open-ended exploration, and strategic decision-making required in realistic pentesting. In this paper, we present a practical evaluation protocol that shifts assessment from task completion to validated vulnerability discovery, allowing evaluation in sufficiently complex targets spanning multiple attack surfaces and vulnerability classes. The protocol combines structured ground-truth with LLM-based semantic matching to identify vulnerabilities, bipartite resolution to score findings under realistic ambiguity, continuous ground-truth maintenance, repeated and cumulative evaluation of stochastic agents, efficiency metrics, and reduced-suite selection for sustainable experimentation. This protocol extends the state of the art by enabling a more realistic, operationally informative comparison of AI pentesting agents. To enable reproducibility, we also release expert-annotated ground truth and code for the proposed evaluation protocol: https://github.com/ethiack/ethibench.

Agent benchmarks increasingly evaluate repository editing, web research, terminal use, and long-horizon interaction. Their scores support capability claims only when the evaluation protocol keeps the intended capability necessary for success. Recent reward-hacking benchmarks and system reports show that agents can instead recover public solutions, read evaluation artifacts, infer generator structure, manipulate feedback, or benefit from invalid scoring paths; existing responses do not provide a common procedure for attributing these shortcuts and quantifying their effect across benchmarks. We formulate protocol validity and introduce HackDetect, a post-hoc audit that identifies an exposure, determines how the agent used it, and assesses whether the resulting score is misleading. We quantify score inflation with the Mislead gap, defined as the exploit score minus the intended score. We audit 2,385 traces across 15 agent benchmarks and find evidence of exposures and reward hacking in 67.0% of Frontier Science traces and 66.7% of AutoLab tasks. Across paired comparisons, we measure score inflation of 0.45-1.00, showing that benchmark reports should provide evidence that scores reflect the intended capability.

As language models (LMs) evolve from chat assistants to long-horizon agents capable of multi-step reasoning and tool use, existing benchmarks remain largely confined to structured or exam-style tasks that fall short of real-world professional demands. To this end, we introduce \OneMillion-Bench OneMillion-Bench, a benchmark of 400 expert-curated tasks spanning Law, Finance, Industry, Healthcare, and Natural Science, built to evaluate agents across economically consequential scenarios. Unlike prior work, the benchmark requires retrieving authoritative sources, resolving conflicting evidence, applying domain-specific rules, and making constraint decisions, where correctness depends as much on the reasoning process as the final answer. We adopt a rubric-based evaluation protocol scoring factual accuracy, logical coherence, practical feasibility, and professional compliance, focused on expert-level problems to ensure meaningful differentiation across agents. Together, \$OneMillion-Bench provides a unified testbed for assessing agentic reliability, professional depth, and practical readiness in domain-intensive scenarios.

Biological protocols are fundamental to reproducible and safe life science research. While LLMs excel on general tasks, their systematic evaluation on these highly specialized, accuracy-critical, and inherently procedural texts remains limited. In this work, we present BioProBench, the first large-scale, integrated multi-task benchmark for biological protocol understanding and reasoning. While limited benchmarks have touched upon specific aspects like protocol QA, BioProBench provides a comprehensive suite of five core tasks: Protocol Question Answering, Step Ordering, Error Correction, Protocol Generation, and Protocol Reasoning, enabling a holistic evaluation of LLMs on procedural biological texts. Built upon 27K original protocols, it yields nearly 556K high-quality structured instances. We evaluate 12 mainstream open/closed-source LLMs on BioProBench. Experimental results reveal that while top models preform well on surface understanding tasks, struggle significantly with deep reasoning and structured generation tasks like ordering and generation. Furthermore, model comparisons reveal diverse performance: certain open-source models approach closed-source levels on some tasks, yet bio-specific small models lag behind general LLMs, indicating limitations on complex procedural content. Overall, our findings underscore that procedural reasoning within biological protocols represents a significant challenge for current LLMs. BioProBench serves as a standardized framework to diagnose these specific limitations and guide the development of AI systems better equipped for safely automating complex scientific procedures. The code and data are available at: https://github.com/YuyangSunshine/bioprotocolbench and https://huggingface.co/datasets/GreatCaptainNemo/BioProBench.

Large language models are increasingly deployed as autonomous agents executing multi-step workflows in real-world software environments. However, existing agent benchmarks suffer from three critical limitations: (1) trajectory-opaque grading that checks only final outputs, (2) underspecified safety and robustness evaluation, and (3) narrow modality coverage and interaction paradigms. We introduce Claw-Eval, an end-to-end evaluation suite addressing all three gaps. It comprises 300 human-verified tasks spanning 9 categories across three groups (general service orchestration, multimodal perception and generation, and multi-turn professional dialogue). Every agent action is recorded through three independent evidence channels (execution traces, audit logs, and environment snapshots), enabling trajectory-aware grading over 2,159 fine-grained rubric items. The scoring protocol evaluates Completion, Safety, and Robustness, reporting Average Score, Pass@k, and Pass^k across three trials to distinguish genuine capability from lucky outcomes. Experiments on 14 frontier models reveal that: (1) trajectory-opaque evaluation is systematically unreliable, missing 44% of safety violations and 13% of robustness failures that our hybrid pipeline catches; (2) controlled error injection primarily degrades consistency rather than peak capability, with Pass^3 dropping up to 24% while Pass@3 remains stable; (3) multimodal performance varies sharply, with most models performing poorer on video than on document or image, and no single model dominating across all modalities. Beyond benchmarking, Claw-Eval highlights actionable directions for agent development, shedding light on what it takes to build agents that are not only capable but reliably deployable.

Evaluating human-AI co-design requires testing whether coding agents can translate real users' multimodal requests into working application revisions. We introduce Co-Design Bench, a reference-free benchmark of 273 tasks curated from 37,463 deduplicated queries across 26,314 real user coding sessions. Its two tracks, Multimodal Editing (ME) and Repository Editing (RE), cover localized edits with optional image, audio, video, or document inputs and coordinated repository revisions. Tasks span three capability dimensions: Grounded Understanding, UI Transformation, and Functional Implementation. The tasks pair normalized requests and runnable pre-edit applications with 2,331 hidden expert-authored acceptance criteria. Our judge-then-refute protocol grades criteria using source and browser evidence, challenges positive verdicts with a separate LLM verifier, and audits visual and interaction claims without gold code or target screenshots. Under a shared agentic harness, the strongest of ten model endpoints passes an average of 69.9% of criteria per task but fully satisfies only 26.4% of tasks. All ten endpoints achieve lower criterion completion on repository edits spanning three or four change types than on those spanning one or two. Co-Design Bench provides a testbed for studying the gap between partial requirement satisfaction and complete task success under agentic execution for real-user application revisions.

This paper reports a single, fully instrumented case study of a large-scale architectural refactoring by an AI coding agent under a specification-first protocol, with no human review of the generated code and no pre-existing oracle to validate the target behaviour. The task, dismantling a central invariant across a large interdependent codebase, was assessed by the author as effectively infeasible through incremental refactoring, the kind of change that conventionally calls for a rewrite instead. Under the protocol described here, the agent completed it successfully. The system is a 717,725-line production TypeScript application across 3,648 files. The task required dismantling a core lifetime invariant: the guarantee that a UI panel remains open for the duration of an AI request. The target behaviour was that a streaming generation survives the closing of its panel and can be reattached, on reopening, to the same live stream with no loss or duplication. The protocol: formal specification by the agent, 14 refinement cycles auditing that specification against the source code, atomic implementation, a compile/test feedback loop, then 17 verification cycles auditing the code against the frozen specification. Across 31 audit passes, 201 defects were corrected before any human executed the program. The convergence criterion was empirical: two consecutive verification passes returning zero findings. The change touched 189 files (31 new); with the extraction phase, the two commits total 288 files, 34,770 insertions, 16,422 deletions. Across the first and roughly thirty later sessions, the software behaved as specified, no bug observed. Elapsed: three days; cost: USD 2,430. The full specification and raw session logs, 1,500+ pages in French, are published as evidence, allowing inspection of the process and submission to a language model for consistency checking.

Systems that generate scientific research questions are evaluated today by expert scores, LLM-as-judge ratings, or curated case studies -- all subjective, none falsifiable. We formalize historical backtesting as an alternative: a system generates questions from a corpus frozen at a historical cutoff, the questions are frozen before any access to later literature, and a temporally isolated future corpus then determines whether each question was subsequently answered, partially addressed, independently posed, or ignored, and whether its underlying premise was supported or refuted. The protocol is model-agnostic: any system that emits frozen questions can be scored. We release reproducible astronomy instances with temporally isolated corpora, frozen questions, auditable labels, four reference baselines, and a submission interface. Two findings result. First, evidence-structure-first generation outperforms LLM-only prompting: across a generator decomposition crossed with a four-cutoff stress test (2010-2024, 798 judged questions) whose last window postdates model training, LLM-only generation shows memorized relevance without specific foresight, while a generator using no model weights at all finds questions whose premises the future refutes in every era. Second, a seven-rater agreement study (two blinded human annotators, five judge models, 90 items) indicts the outcome taxonomy rather than the judge: two careful humans agree at kappa = 0.17, every judge model agrees with the professional annotator as well or better (0.17-0.26), and frontier models agree with one another at 0.60 -- certifying an LLM judge by model-model agreement would have overstated its reliability threefold. A prospective instance -- 200 questions frozen 2026-08-17, scored 2027-2030 -- is released so the central claims become contamination-free tests that time itself will grade.

Recent advances have made long-form report-generating systems widely available. This has prompted evaluation frameworks that use LLM-as-judge protocols and claim verification, along with meta-evaluation frameworks that seek to validate these methods. Many of the meta-evaluations estimate an evaluation quality's by comparing its assessments against human pairwise preferences. Prior work, however, suggests that human pairwise preference may be overly simplistic and can fail to capture nuances of expert expectations. We conduct a case study in meta-evaluation for long-form QA benchmarks using ScholarQA-CS2, a benchmark designed for assessing retrieval-augmented deep-research QA in the scientific domain. We comprehensively validate the benchmark through human pairwise preference judgments, then critically examine the strengths, weaknesses, and confounders of this approach. We show that pairwise preference rankings are best suited for system-level evaluation, while explicit metric-wise annotations and expert annotators are critical for reliable metric-level assessment, with subjectivity remaining a key challenge. Based on our findings, we offer practical guidelines for designing future meta-evaluations that better align evaluation methods, annotator expertise, and reporting practices. By surfacing these methodological challenges, we aim to advance evaluation standards for deep-research systems.

A coding agent must emit a valid tool call--a parseable invocation of a tool in the provided schema--before the harness can execute its chosen action. We study how local serving stacks affect this protocol step and show that measured outcomes can depend on the serving layer rather than model behavior alone. In Ollama, the default tools= request is gated per model by a static template flag: some models are accepted and return calls as text, some return native tool_calls, while Phi-3 and Gemma-3 are rejected before inference. In our harness, rejection and retry exhaustion are not preserved as structured failure metadata, so downstream analysis can misclassify them as model non-calls and naively report 0% fidelity. Adding a text tool list while retaining the native channel recovers much of the measured fidelity for accepted models, whereas a uniform text protocol reduces fidelity for Llama-3.2, which has native tool-call support. Cross-stack probes on Ollama, llama.cpp, vLLM, and SGLang show different handling of the same request. Constrained decoding removes parse failures but can induce non-termination, and turn-pooled versus per-instance estimates differ by up to about 55 points. We conclude with a checklist for treating serving behavior as part of the evaluation protocol.

The critique capacity of Large Language Models (LLMs) is essential for reasoning abilities, which can provide necessary suggestions (e.g., detailed analysis and constructive feedback). Therefore, how to evaluate the critique capacity of LLMs has drawn great attention and several critique benchmarks have been proposed. However, existing critique benchmarks usually have the following limitations: (1). Focusing on diverse reasoning tasks in general domains and insufficient evaluation on code tasks (e.g., only covering code generation task), where the difficulty of queries is relatively easy (e.g., the code queries of CriticBench are from Humaneval and MBPP). (2). Lacking comprehensive evaluation from different dimensions. To address these limitations, we introduce a holistic code critique benchmark for LLMs called CodeCriticBench. Specifically, our CodeCriticBench includes two mainstream code tasks (i.e., code generation and code QA) with different difficulties. Besides, the evaluation protocols include basic critique evaluation and advanced critique evaluation for different characteristics, where fine-grained evaluation checklists are well-designed for advanced settings. Finally, we conduct extensive experimental results of existing LLMs, which show the effectiveness of CodeCriticBench.

LeWorldModel trains a latent world model with a prediction loss and a single anti-collapse regulariser, and reports approximately 87% of goals reached on TwoRoom, its simplest diagnostic environment. We reproduce that result by independent reimplementation on roughly $25 of rented compute, with all evaluation on one laptop CPU. We reach 94.0% at the repository's evaluation goal offset, against 84.0% for the authors' own released checkpoint measured under our protocol on identical episodes, and we reproduce the reported representation result directly (position probe Pearson r = 0.9988 against a reported 0.996). Reaching that point required four conventions that determine the outcome and appear in no released configuration file: dense action gathering across a frameskip block, a programmatically-set action-encoder width, ImageNet pixel normalisation, and action z-scoring. A reproducer following the released configurations alone obtains a model whose predictor cannot converge. The evaluation protocol is itself contested by the released material. The paper's appendix and the repository's configuration specify different goal offsets and step budgets; on the authors' own weights these yield 14.0% and 84.0%, and only the configuration's values reproduce the reported figure. On fifty identical episodes, changing nothing but how the goal is constructed moves that checkpoint from 84.0% to 8.0%. Two findings generalise. One-step prediction accuracy does not predict long-horizon planning success: across three checkpoints spanning a sevenfold range in prediction error, including the authors' own, it orders short-horizon success monotonically and fails to order long-horizon success at all. And a batch normalisation layer inflated our reported validation loss by up to a factor of 300, concealing a training loss that was flat throughout.

Rigorous and reproducible evaluation is critical for assessing the state of the art and for guiding scientific advances in Artificial Intelligence. Evaluation is challenging in practice due to several reasons, including benchmark saturation, lack of transparency in methods used for measurement, development challenges in extracting measurements for generative tasks, and, more generally, the extensive number of capabilities required for a well-rounded comparison across models. We make three contributions to alleviate the above challenges. First, we present Eureka, an open-source framework for standardizing evaluations of large foundation models beyond single-score reporting and rankings. Second, we introduce Eureka-Bench as an extensible collection of benchmarks testing capabilities that (i) are still challenging for state-of-the-art models and (ii) represent fundamental but overlooked language and multimodal capabilities. The inherent space for improvement in non-saturated benchmarks enables us to discover meaningful differences between models at a capability level. Third, using Eureka, we conduct an analysis of 12 state-of-the-art models, providing in-depth insights into failure understanding and model comparison, which can be leveraged to plan targeted improvements. In contrast to recent trends in reports and leaderboards showing absolute rankings and claims for one model or another to be the best, our analysis shows that there is no such best model. Different models have different strengths, but there are models that appear more often than others as best performers for some capabilities. Despite the recent improvements, current models still struggle with several fundamental capabilities including detailed image understanding, benefiting from multimodal input when available rather than fully relying on language, factuality and grounding for information retrieval, and over refusals.

Benchmarks are essential for quantitatively tracking progress in AI. As AI agents become increasingly capable, researchers and practitioners have introduced agentic benchmarks to evaluate agents on complex, real-world tasks. These benchmarks typically measure agent capabilities by evaluating task outcomes via specific reward designs. However, we show that many agentic benchmarks have issues task setup or reward design. For example, SWE-bench Verified uses insufficient test cases, while TAU-bench counts empty responses as successful. Such issues can lead to under- or overestimation agents' performance by up to 100% in relative terms. To make agentic evaluation rigorous, we introduce the Agentic Benchmark Checklist (ABC), a set of guidelines that we synthesized from our benchmark-building experience, a survey of best practices, and previously reported issues. When applied to CVE-Bench, a benchmark with a particularly complex evaluation design, ABC reduces the performance overestimation by 33%.

The computation necessary for training Transformer-based language models has skyrocketed in recent years. This trend has motivated research on efficient training algorithms designed to improve training, validation, and downstream performance faster than standard training. In this work, we revisit three categories of such algorithms: dynamic architectures (layer stacking, layer dropping), batch selection (selective backprop, RHO loss), and efficient optimizers (Lion, Sophia). When pre-training BERT and T5 with a fixed computation budget using such methods, we find that their training, validation, and downstream gains vanish compared to a baseline with a fully-decayed learning rate. We define an evaluation protocol that enables computation to be done on arbitrary machines by mapping all computation time to a reference machine which we call reference system time. We discuss the limitations of our proposed protocol and release our code to encourage rigorous research in efficient training procedures: https://github.com/JeanKaddour/NoTrainNoGain.

We introduce MacroBench, a code-first benchmark that evaluates whether LLMs can synthesize reusable browser-automation programs (macros) from natural-language goals by reading HTML/DOM and emitting Selenium. MacroBench instantiates seven self-hosted sites covering 681 tasks across interaction complexity and targeting difficulty. Our end-to-end protocol validates generated code via static checks, sandboxed execution, and outcome verification (DOM assertions, database snapshots), and includes a safety suite for scraping, spam/abuse, and credential/privacy prompts. Across 2,636 model-task runs, we observe stratified success: GPT-4o-mini (96.8%), GPT-4o (95.3%), Gemini (89.0%), DeepSeek (83.4%). Models handle simple tasks reliably (91.7%) but fail on complex workflows (0.0%), and none meet production-quality coding practices despite functional completion. We release our complete benchmark pipeline, evaluation framework, and experimental results at https://github.com/hyunjun1121/MacroBench to enable reproducible assessment of macro synthesis for web automation.

AI research agents combine prior knowledge, public sources, and experimental feedback to produce useful results. The Discovery Certification Protocol (DCP) turns claims about these results into executable recovery and feedback tests. Gate 1 validates useful improvement on sealed evaluation. Gate 2 gives matched agents the registered starting information and observed Web content while withholding the target research history. Every valid method reaching the numerical target supplies a recovery witness and triggers the Core veto. DCP Core requires adequate controls, zero observed recoveries, and a finite-sample bound on recovery in one fresh registered episode. Optional Gate 3 measures the average effect of truthful feedback relative to a specified neutral policy from a shared checkpoint. DCP Evidence adds this effect after independent null calibration and a registered effect margin. Two controlled audits exercise the complete protocol in SQLite optimization and virtual catalyst control under different models. Each produced zero recoveries in 96 episodes, with an upper bound of 0.0468. Each paired study yielded 30 truthful recoveries and zero neutral recoveries, with passing 60-pair null studies. Additional cases exercise Core, recovered, and audit-incomplete decisions. A deterministic, LLM-free verifier reproduces the decisions from frozen evidence. DCP provides a common evidence language for useful outcomes, alternative routes, and feedback effects across AI research.

Runtime-governance evidence often collapses materially different events into one audit record: a decision may be made, an instruction dispatched or received, an action may or may not execute, and a resulting state may or may not be observed. This paper presents AIREP, a vendor- and model-independent protocol for per-decision AI runtime evidence. Its current wire model separates evidence into four artifact families: Decision, Control, Execution, and Effect. Artifacts use closed core schemas, explicit identities and digests, declared scope limits, RFC 8785 canonical JSON, domain-separated SHA-256 hashing, and pure Ed25519 signatures. A three-level assurance model distinguishes structural/hash consistency (AIREP-Core), verifier-accepted authorship (AIREP-Authenticated), and independently anchored chain-head freshness and non-truncation relative to an accepted witness (AIREP-Witnessed); these classes do not establish event truth. A structured reconciler preserves failure, missing evidence, unevaluated prerequisites, and indeterminate outcomes as distinct states. The released beta includes a four-family first-party producer, Python and Node reference-verification paths, adversarial/lifecycle corpora, and reproducible validation. Post-release first-party Hermes and LightEval integration exercises preserve explicit evidence boundaries without claiming adoption or interoperability. Independent implementation evidence exists separately for a v0.1.2 producer and a v0.2 consumer/verifier; because they target different frozen versions, they do not establish same-version producer-to-consumer interoperability. AIREP remains experimental.

Scientific images function as critical evidence in research communication, yet their integrity faces unprecedented threats from AI-generated content that introduces subtle but consequential errors. Existing evaluation paradigms prove inadequate: perceptual quality metrics poorly correlate with scientific validity, while language models lack domain-specific verification capabilities. To address this gap, we propose the Scientific Image Utility and Upgradability Assessment (SIU^2A) framework, which introduces two complementary dimensions for scientific image evaluation. Utility encompasses error detection (identifying scientific inaccuracies) and correction feasibility (assessing whether errors can be reliably repaired). Upgradability measures the quality of correction. We categorize scientific image corruption into four fundamental types: Detail Distortion, Incompleteness, False Content, and Entity Confusion. Based on this taxonomy, we construct SIU^2A-Benchmark, a dataset with expert annotations for error identification and repair. The framework implements a two-stage evaluation protocol: the Utility stage evaluates error detection capability and repair instruction generation, while the Upgradability stage assesses whether corrections faithfully restore scientific validity without compromising existing accurate information. Experiments reveal that current multimodal systems exhibit significant limitations in both scientific error assessment and faithful correction, exposing a fundamental gap between visual perception and scientific usability.

Personal agents maintain memories, learned skills, tool configurations, and policy state that evolve with each user. Existing agent benchmarks often evaluate these capabilities in isolation: tool benchmarks test invocation under fixed APIs, memory benchmarks test recall or forgetting, and safety benchmarks test static policy compliance. We argue that personal-agent evaluation requires a different protocol: replaying the same temporal intervention across different persistent user-conditioned states and measuring how failures propagate across agent components. We formalize this requirement as four conditions: explicit temporal intervention, persistent state across the intervention, induced cross-dimensional effects, and variation in user-conditioned state. A focused audit of public benchmark protocols selected by explicit inclusion criteria identifies several close cases. Under our explicitly narrow operationalization, we did not find a protocol in that audited set satisfying all four conditions. This claim is scoped as a focused gap analysis with bounded literature coverage. This position paper proposes a minimal benchmark design and candidate reporting metrics for user-conditioned adaptation. The result is a concrete design requirement for future personal-agent evaluation, with metrics used as reporting tools for that requirement.

Various algorithms for continual learning (CL) have been designed with the goal of effectively alleviating the trade-off between stability and plasticity during the CL process. To achieve this goal, tuning appropriate hyperparameters for each algorithm is essential. As an evaluation protocol, it has been common practice to train a CL algorithm using diverse hyperparameter values on a CL scenario constructed with a benchmark dataset. Subsequently, the best performance attained with the optimal hyperparameter value serves as the criterion for evaluating the CL algorithm. In this paper, we contend that this evaluation protocol is not only impractical but also incapable of effectively assessing the CL capability of a CL algorithm. Returning to the fundamental principles of model evaluation in machine learning, we propose an evaluation protocol that involves Hyperparameter Tuning and Evaluation phases. Those phases consist of different datasets but share the same CL scenario. In the Hyperparameter Tuning phase, each algorithm is iteratively trained with different hyperparameter values to find the optimal hyperparameter values. Subsequently, in the Evaluation phase, the optimal hyperparameter values is directly applied for training each algorithm, and their performance in the Evaluation phase serves as the criterion for evaluating them. Through experiments on CIFAR-100 and ImageNet-100 based on the proposed protocol in class-incremental learning, we not only observed that the existing evaluation method fail to properly assess the CL capability of each algorithm but also observe that some recently proposed state-of-the-art algorithms, which reported superior performance, actually exhibit inferior performance compared to the previous algorithm.

Unsupervised physical parameter estimation from video lacks a common benchmark: existing methods evaluate on non-overlapping synthetic data, the sole real-world dataset is restricted to single-body systems, and no established protocol addresses governing-equation identification. This work introduces IRIS, a high-fidelity benchmark comprising 240 real-world videos captured at 4K resolution and 60fps, spanning both single- and multi-body dynamics with independently measured ground-truth parameters and uncertainty estimates. Each dynamical system is recorded under controlled laboratory conditions and paired with its governing equations, enabling principled evaluation. A standardized evaluation protocol is defined encompassing parameter accuracy, identifiability, extrapolation, robustness, and governing-equation selection. Multiple baselines are evaluated, including a multi-step physics loss formulation and four complementary equation-identification strategies (VLM temporal reasoning, describe-then-classify prompting, CNN-based classification, and path-based labelling), establishing reference performance across all IRIS scenarios and exposing systematic failure modes that motivate future research. The dataset, annotations, evaluation toolkit, and all baseline implementations are publicly released.

Many math- and science-oriented agent systems use hierarchical designs with specialized reviewer roles, assuming that a dedicated review stage should help turn wrong candidates into correct ones. We test this assumption on 4,181 verifier-grounded Omni-MATH problems using matched gpt-oss-120b actors. Collaboration adds little on the easiest tiers, but from tier 4 onward the gains open sharply; in this harder regime, broadcast-style peer discussion reaches higher final accuracy than a planner-executor-reviewer pipeline (PER). We ask whether this gap is explained by reviewer quality or by whether critique changes the next answer the protocol carries forward. It is not explained by reviewer precision alone: PER's reviewer is more precise than broadcast's (0.861 vs. 0.644), yet evaluator-verified useful critique is much less likely to change the next candidate and produces lower reviewer-guided repair. These results show that reviewer detection quality and critique uptake are empirically separable. Within matched PER interventions, forcing explicit acknowledgment lowers final accuracy, while embedding reviewer guidance directly in the solver's working context partially improves follow-through without closing the gap. Overall, reviewer-centric evaluation can overstate system quality: a protocol may spot errors well yet still fail to solve more problems if it does not act on those critiques.

An AI control protocol is a plan for usefully deploying AI systems that aims to prevent an AI from intentionally causing some unacceptable outcome. This paper investigates how well AI systems can generate and act on their own strategies for subverting control protocols whilst operating statelessly (without shared memory between contexts). To do this, an AI system may need to reliably generate optimal plans in each context, take actions with well-calibrated probabilities, and coordinate plans with other instances of itself without communicating. We develop Subversion Strategy Eval, a suite of eight environments, covering a range of protocols and strategic capabilities, and six sets of affordances that help isolate individual capabilities. We implement the evaluation in Inspect-AI and release it open-source. We evaluate Claude 3.5 models, including helpful-only versions, as well as OpenAI reasoning models. None of the models demonstrate substantial capability in strategizing to subvert control protocols statelessly. However, providing models with additional affordances, such as the ability to share a plan between contexts, can substantially improve performance. We hope our evaluations can act as a leading indicator for when models are capable of subverting control protocols and also relax the worst-case assumption of perfect strategic ability in AI control evaluations.

Recent machine learning papers often report 1-2 percentage point improvements from a single run on a benchmark. These gains are highly sensitive to random seeds, data ordering, and implementation details, yet are rarely accompanied by uncertainty estimates or significance tests. It is therefore unclear when a reported +1-2% reflects a real algorithmic advance versus noise. We revisit this problem under realistic compute budgets, where only a few runs are affordable. We propose a simple, PC-friendly evaluation protocol based on paired multi-seed runs, bias-corrected and accelerated (BCa) bootstrap confidence intervals, and a sign-flip permutation test on per-seed deltas. The protocol is intentionally conservative and is meant as a guardrail against over-claiming. We instantiate it on CIFAR-10, CIFAR-10N, and AG News using synthetic no-improvement, small-gain, and medium-gain scenarios. Single runs and unpaired t-tests often suggest significant gains for 0.6-2.0 point improvements, especially on text. With only three seeds, our paired protocol never declares significance in these settings. We argue that such conservative evaluation is a safer default for small gains under tight budgets.

State machine replication (SMR) and Byzantine fault-tolerant (BFT) consensus guarantee agreement despite a bounded number of arbitrary, colluding faulty participants. However, these guarantees rely on participants outside this set correctly executing the protocol's transition semantics. Agentic validators expose a weaker boundary: an authenticated, responsive, non-equivocating, and protocol-compliant reasoning participant may still endorse a semantically invalid transition due to reasoning errors. We call this failure mode an epistemic fault, and the collective phenomenon the Honest Quorum Problem (where "honest" means protocol-compliant, not semantically correct). Such a quorum can satisfy ordinary checks while forming a certificate for an invalid transition. Thus, agreement alone does not guarantee semantic validity or execution safety. Furthermore, because agentic validators often share model weights, training distributions, prompts, or toolchains, they are highly susceptible to correlated epistemic faults. We define Epistemic Byzantine Fault Tolerance (EBFT), a fault-tolerance model for agentic infrastructure and post-deterministic distributed systems. EBFT augments the conventional Byzantine fault bound with two separate, confidence-indexed quantities: e_δ bounds coherent invalid endorsements outside the Byzantine set, and u_ε bounds unusable validator support that degrades liveness. These quantities characterize semantic safety risk and liveness degradation independently. We derive quorum-threshold conditions for semantic validity, consensus agreement, liveness, and feasible threshold selection, and outline a calibration methodology for estimating these budgets. We show that adding nominally distinct agents improves fault tolerance only when it measurably reduces the upper-tail concentration of invalid endorsements or unusable support.

Large language models are increasingly capable of synthesizing executable frontend projects, yet existing benchmarks still treat web generation as a static evaluation problem. We argue that frontend artifacts demand a different paradigm: they are interactive rather than static, admit diverse yet equally valid implementations, and evolve faster than rigid pipelines can accommodate. To address these gaps, we present LiveEvalBench, an automated framework that reformulates web-generation evaluation as an agentic, adaptive, and extensible process. LiveEvalBench instantiates evaluation as a collaborative review workflow, in which a Build Engineer, a Code Engineer, and a UI Tester collectively gather evidence across the full lifecycle of a frontend project, from deployment and code inspection to browser-based interaction. To handle implementation diversity, an adaptive protocol couples shared rubrics for cross-model comparability with implementation-grounded criteria tailored to each artifact. The framework further supports incremental integration of new evaluator roles and assessment dimensions without pipeline redesign. Experiments across diverse real-world web-generation scenarios show that LiveEvalBench aligns closely with human expert judgment and provides fine-grained insights into frontier models' web generation capabilities. Code is available at https://github.com/wyysteelhead/LiveEvalBench

The Model Context Protocol (MCP) defines a schema bound execution model for agent-tool interaction, enabling modular computer vision workflows without retraining. To our knowledge, this is the first protocol level, deployment scale audit of MCP in vision systems, identifying systemic weaknesses in schema semantics, interoperability, and runtime coordination. We analyze 91 publicly registered vision centric MCP servers, annotated along nine dimensions of compositional fidelity, and develop an executable benchmark with validators to detect and categorize protocol violations. The audit reveals high prevalence of schema format divergence, missing runtime schema validation, undeclared coordinate conventions, and reliance on untracked bridging scripts. Validator based testing quantifies these failures, with schema format checks flagging misalignments in 78.0 percent of systems, coordinate convention checks detecting spatial reference errors in 24.6 percent, and memory scope checks issuing an average of 33.8 warnings per 100 executions. Security probes show that dynamic and multi agent workflows exhibit elevated risks of privilege escalation and untyped tool connections. The proposed benchmark and validator suite, implemented in a controlled testbed and to be released on GitHub, establishes a reproducible framework for measuring and improving the reliability and security of compositional vision workflows.

Large language models (LLMs) deployed for logical reasoning in knowledge-intensive domains exhibit a subtle but critical failure: coherence can be vacuously achieved through systematic abstention. A model that withholds commitment to either entailment or refutation satisfies negation consistency while providing no utility. We introduce Coherence Under Commitment (CUC), a dual-query evaluation paradigm that jointly measures consistency and decisiveness. CUC contributes three innovations: (1) a commitment score c(φ) = p(φ) + p(lnotφ) quantifying probability mass allocated to decisive outcomes; (2) a deterministic elicitation protocol via normalized YES/NO log probabilities, eliminating sampling variance; and (3) a 3-way decision framework (True/False/Uncertain) operationalizing the coherence-commitment trade-off into metrics. Experiments on four open-weight LLMs (1B-3B) across 204 FOLIO examples expose a sharp frontier. Qwen2.5-3B achieves near-zero contradiction (E[v_{neg}]{=}0.025) but only 7.4% coverage, while TinyLlama-1.1B reaches 79.4% coverage with violations on every example. Coherence-only evaluation would rank the abstaining model first; CUC exposes this as vacuous, and the frontier generalizes to LogiQA~v2 (ρ{=}0.97). We argue that evaluation must report both coherence and non-vacuous commitment and release a toolkit for standardized assessment.

Corruption studies, the primary tool for evaluating chain-of-thought (CoT) faithfulness, identify which chain positions are "computationally important" by measuring accuracy when steps are replaced with errors. We identify a systematic confound: for chains with explicit terminal answer statements, the dominant format in standard benchmarks, corruption studies detect where the answer text appears, not where computation occurs. A within-dataset format ablation provides the key evidence: on standard GSM8K chains ending with "the answer is X," removing only the answer statement, preserving all reasoning, collapses suffix sensitivity ~19x at 3B (N=300, p=0.022). Conflicting-answer experiments quantify the causal mechanism: at 7B, CC accuracy drops to near-zero (<=0.02) across five architecture families; the followed-wrong rate spans 0.63-1.00 at 3B-7B and attenuates at larger scales (0.300 at Phi-4-14B, ~0.01 at 32B). A within-stable 7B replication (9.3x attenuation, N=76, p=7.8e-3; Qwen3-8B N=299, p=0.004) provides converging evidence, and the pattern replicates on MATH (DeepSeek-R1-7B: 10.9x suffix-survival recovery). On chains without answer suffixes the same protocol identifies the prefix as load-bearing (Delta=-0.77, p<10^-12). Generation-time probes confirm a dissociation: the answer is not early-determined during generation (early commitment <5%), yet at consumption time model outputs systematically follow the explicit answer text. The format-determination effect persists through 14B (8.5x ratio, p=0.001) and converges toward zero at 32B. We propose a three-prerequisite protocol (question-only control, format characterization, all-position sweep) as a minimum standard for corruption-based faithfulness studies.

Always-on agents are systems whose future behavior depends on durable state accumulated across earlier interactions. We treat them as persistent-state systems: the operative system includes retrievable memories, but also task ledgers, permissions, credentials, commitments, provenance and audit records, shared state, trigger conditions, and externally committed effects linked to those records. The survey reads the literature through six diagnostic axes for each state item, authority, scope, mutability, provenance, recoverability, and actionability, and through a lifecycle in which state is written, validated, organized, retrieved, acted upon, updated, forgotten, audited, and sometimes rolled back. Across a 435-work coded corpus, treated as a scoped map rather than an exhaustive census, the literature concentrates more heavily on accumulating and retrieving state than on governing, recovering, or relinquishing it. We therefore introduce the Always-On Evaluation Protocol (AOEP-v0), a pilot evaluation contract that makes these governance requirements concrete by scoring state mutation and recovery obligations rather than answer quality alone. The resulting agenda connects always-on agents to databases, distributed systems, formal methods, capability security, and machine unlearning.

Reality may establish that an outcome occurred without identifying which evolving procedure produced it or why. This distinction matters in production ML systems whose code, configuration, and artifacts change while external feedback accumulates. We examine it in a human-directed, agent-engineered quantitative trading system, using oracle to mean an external source of realized outcomes rather than a complete correctness specification. Across one year, the account gained and outperformed a broad market index, while annual alpha was not statistically distinguishable from zero under the main retrospective specification. Retrospectively selected subperiods include adverse relative performance and conditional candidate-level weakness under declared approximate references. Engineering records document changes during the episode, and complete recommendation-to-runtime binding is unavailable. The archive does not establish a common frozen instance or a unique cause. The case motivates an outcome--diagnosis gap: outcome evidence, evaluated-object identity, and causal explanation support distinct claims. We distinguish frozen instances, pre-specified adaptive procedures, and ad-hoc development; organize archive-relative claim identifiability and an evidence hierarchy; and propose a prospective production-binding protocol. An illustrative compatible-history example shows how factual binding can resolve a recommendation's referent without supplying its counterfactual effect. The protocol is proposed rather than prospectively validated. External feedback constrains outcome claims, while provenance and additional identification structure determine the resolution of diagnosis.

Access to the right evidence does not guarantee that large language models (LLMs) will reason with it correctly. This gap between retrieval and reasoning is especially concerning in clinical settings, where outputs must align with structured protocols. We study this gap using Written Exposure Therapy (WET) guidelines as a testbed. In evaluating model responses to curated clinician-vetted questions, we find that errors persist even when authoritative passages are provided. To address this, we propose an evaluation framework that measures accuracy, consistency, and fidelity of reasoning. Our results highlight both the potential and the risks: retrieval-augmented generation (RAG) can constrain outputs, but safe deployment requires assessing reasoning as rigorously as retrieval.

This position paper argues that computer science conferences should require tamper-evident, nonrepudiable attestations of experimental results. We name the underlying problem experiment nonrepudiation: a compliant protocol must bind the numbers in a paper to an actual executed computation in a way the author cannot later alter or deny. The current system relies on self-reported checklists, optional code sharing, and author-controlled logging. None of these mechanisms answer the question a reviewer cannot check: did the code the paper describes produce the numbers the paper reports? We define the problem formally, state the security properties any compliant protocol must satisfy, and describe a threat model that includes attacks current approaches do not prevent. To show that the problem is solvable, we built K-Veritas, a reference implementation in Go that produces signed reports without accessing training data. K-Veritas is a testbed, not a finished answer. We call on conferences and the community to treat nonrepudiation as a first-class requirement and to help build an open, independent standard for it.

Diagnosing 802.11 packet captures requires expert protocol knowledge, is slow, inconsistent across engineers, and unscalable. LLM-based approaches sound plausible but fabricate protocol events absent from captures (especially truncated traces), produce uncalibrated confidence scores, and suffer evaluation bias when golden references are co-produced by the model under test. We introduce PROBE (Protocol Reasoning Over evidence-Based Ensembles), a multi-stage pipeline addressing all three failures. It integrates (i) deterministic PCAP-to-text normalization with frame-level verifiability, (ii) multi-run, multi-candidate ensembles with optional cross-model second opinion and progressive obfuscation, (iii) a verdict-aware evidence framework treating absence of failure evidence as contributing evidence, and (iv) a fully deterministic composite reliability score from evidence validity, run-to-run stability, and cross-model agreement without LLM self-assessment. On 87 enterprise Wi-Fi captures (104 capture-reviewer pairs), single-pass LLM analysis raises weighted evidence F1 from 0.871 (expert baseline) to 0.912 but misses critical frames in 35% of cases. Naive ensemble voting drops below baseline (0.842) as majority voting amplifies conservative verdicts: 50% of confirmed failures are misclassified as 'no issue' or 'insufficient evidence.' Adding evidence-grounded reconciliation achieves 0.957 F1, a 96% auto-accept rate, and a worst-case floor above 0.70. LLM self-reported confidence clusters at 0.95 regardless of difficulty (71% report exactly 0.95), confirming it is uninformative. We also introduce a model-agnostic evaluation framework using per-field assertion matching, eliminating circular bias from model-co-produced golden references.

The convergence of LLM-powered research assistants and AI-based peer review systems creates a critical vulnerability: fully automated publication loops where AI-generated research is evaluated by AI reviewers without human oversight. We investigate this through BadScientist, a framework that evaluates whether fabrication-oriented paper generation agents can deceive multi-model LLM review systems. Our generator employs presentation-manipulation strategies requiring no real experiments. We develop a rigorous evaluation framework with formal error guarantees (concentration bounds and calibration analysis), calibrated on real data. Our results reveal systematic vulnerabilities: fabricated papers achieve acceptance rates up to . Critically, we identify concern-acceptance conflict -- reviewers frequently flag integrity issues yet assign acceptance-level scores. Our mitigation strategies show only marginal improvements, with detection accuracy barely exceeding random chance. Despite provably sound aggregation mathematics, integrity checking systematically fails, exposing fundamental limitations in current AI-driven review systems and underscoring the urgent need for defense-in-depth safeguards in scientific publishing.

Tool-augmented LLM agents call APIs whose intermediate outputs, such as presigned URLs, session tokens, and OAuth state parameters, are observation contracts: artifacts whose later use is constrained by the external system that produced them. We show that observation contract compliance (preserving the temporal validity and byte-level integrity) is an emergent, regression-prone capability: it is neither guaranteed by general tool-use ability nor consistently improved by larger or newer models. To measure this, we introduce ContractBench, a benchmark of 33 dual-axis tasks that probe two orthogonal failure modes no existing benchmark evaluates: validity failures (using an artifact after expiry) and integrity failures (corrupting an artifact's bytes through the observation-to-action pipeline). Our evaluation is deterministic and programmatic, with a virtual clock controlling time and SHA-256 hashes verifying byte integrity. We assign each outcome a failure label drawn from real-world API specifications. We evaluate 38 models and report four findings: (i) no evaluated model clears 80%, with Claude-Opus-4.6 leading at 77.8%, revealing that current frontier models still fail to comply with observation contracts; (ii) a sharp within-family capability cliff in Qwen 3.5 between 4B (0%) and 9B (56.6%), smoothing to 70.7% at 397B-A17B: what emerges across the cliff is mid-trajectory restraint, not tool-call competence; (iii) non-monotonic scaling across the GPT-5 family: agentic post-training can erode compliance through sycophancy-driven regression; (iv) our failure taxonomy works as an actionable in-context reward signal, yielding +7.1 pp on 42 paired GPT-5.1 failures.

Large language models (LLMs) augmented with tools are emerging as autonomous agents capable of using Web engine, APIs, and code to solve complex, long-horizon tasks. Current tool-using benchmarks for information seeking on academic graphs rely on synthetic templates, simplified solution spaces, or narrow tasks such as paper-centric tasks, leaving key challenges underexplored - realistic user intent, complex multi-step API planning, rich parameter filling for APIs, grounded answers with references, and comprehensive evaluation of both the process and the outcome. We introduce AISE-Bench, a real-world, full-cycle annotated benchmark for information seeking on academic knowledge graphs. AISE-Bench release contains 1,133 QA pairs, including query taxonomies, full API execution trajectories, validated parameters, and source-grounded answers with reference links. To support high-quality annotation, we design a customized agent workflow to enable annotators to plan, execute, and revise complex API workflows efficiently. We develop a comprehensive evaluation protocol measuring answer quality, reference grounding, API-planning correctness, and execution success. Among the 14 evaluated methods, even the strongest model (PLAY2PROMPT with Gemini-3-Pro) achieves only moderate performance and often struggles with API planning and execution. AISE-Bench establishes a challenging new testbed for quantitatively evaluating and improving the stepwise correctness, grounded summarization, and traceable reasoning of multi-step API-using LLM agents. Our code and data are available at https://aise-bench.github.io/.

Video summarization is a technique to create a short skim of the original video while preserving the main stories/content. There exists a substantial interest in automatizing this process due to the rapid growth of the available material. The recent progress has been facilitated by public benchmark datasets, which enable easy and fair comparison of methods. Currently the established evaluation protocol is to compare the generated summary with respect to a set of reference summaries provided by the dataset. In this paper, we will provide in-depth assessment of this pipeline using two popular benchmark datasets. Surprisingly, we observe that randomly generated summaries achieve comparable or better performance to the state-of-the-art. In some cases, the random summaries outperform even the human generated summaries in leave-one-out experiments. Moreover, it turns out that the video segmentation, which is often considered as a fixed pre-processing method, has the most significant impact on the performance measure. Based on our observations, we propose alternative approaches for assessing the importance scores as well as an intuitive visualization of correlation between the estimated scoring and human annotations.

Standard evaluation protocols reveal a counterintuitive phenomenon: on 7.7% of benchmark problems spanning five datasets, larger language models underperform smaller ones by 28.4 percentage points despite 10-100x more parameters. Through systematic evaluation of 31 models (0.5B-405B parameters) across 1,485 problems, we identify the mechanism as spontaneous scale-dependent verbosity that introduces errors through overelaboration. Causal intervention experiments demonstrate this reflects correctable prompt design rather than fundamental capability limitations. Constraining large models to produce brief responses improves accuracy by 26 percentage points and reduces performance gaps by up to two-thirds. Most critically, brevity constraints completely reverse performance hierarchies on mathematical reasoning and scientific knowledge benchmarks, with large models achieving 7.7-15.9 percentage point advantages over small models -- direct inversions of the original gaps. These reversals prove large models possess superior latent capabilities that universal prompting masks. We validate findings through three independent contamination tests and demonstrate inverse scaling operates continuously across the full parameter spectrum, with dataset-specific optimal scales ranging from 0.5B to 3.0B parameters. Our results establish that maximizing large model performance requires scale-aware prompt engineering rather than universal evaluation protocols, with immediate implications for deployment: prompt adaptation simultaneously improves accuracy and reduces computational costs.

In this study, we introduce FEET, a standardized protocol designed to guide the development and benchmarking of foundation models. While numerous benchmark datasets exist for evaluating these models, we propose a structured evaluation protocol across three distinct scenarios to gain a comprehensive understanding of their practical performance. We define three primary use cases: frozen embeddings, few-shot embeddings, and fully fine-tuned embeddings. Each scenario is detailed and illustrated through two case studies: one in sentiment analysis and another in the medical domain, demonstrating how these evaluations provide a thorough assessment of foundation models' effectiveness in research applications. We recommend this protocol as a standard for future research aimed at advancing representation learning models.

Human evaluation is critical for validating the performance of text-to-image generative models, as this highly cognitive process requires deep comprehension of text and images. However, our survey of 37 recent papers reveals that many works rely solely on automatic measures (e.g., FID) or perform poorly described human evaluations that are not reliable or repeatable. This paper proposes a standardized and well-defined human evaluation protocol to facilitate verifiable and reproducible human evaluation in future works. In our pilot data collection, we experimentally show that the current automatic measures are incompatible with human perception in evaluating the performance of the text-to-image generation results. Furthermore, we provide insights for designing human evaluation experiments reliably and conclusively. Finally, we make several resources publicly available to the community to facilitate easy and fast implementations.

Standard accuracy benchmarks are designed to test how closely large language models (LLMs) approach correct answers, but are not suitable for testing whether LLMs stick with a correct answer when that answer is challenged by a plausible counter-argument. We introduce a controlled protocol for evaluating answer stability: after a model answers a multiple-choice question correctly, we challenge the model's answer with a coherent argument for an incorrect option and measure whether the model flips. The setup a) isolates argumentative content from overt social pressure and b) varies argument length, self-attribution, and cross-model source. Across seven frontier models and 57 MMLU subjects, flip rates range from 17.5% to 97.3%, revealing large differences in stability that are not captured by accuracy metrics alone. We find that self-attribution consistently increases flip rates (mean +7.1pp, up to +18.7pp). Also, pooling wrong-answer arguments across models and selecting the most effective one per question yields stronger adversarial challenges than relying on any single source model. We further construct MaxFlip, a curated challenge set that amplifies flips by up to +23.6pp over standard self-generated challenges. We release the protocol, challenge records, and MaxFlip to support stability evaluation alongside standard accuracy benchmarks. Materials are available at https://github.com/nafisenik/WhoFlips and https://hf.co/datasets/nafisehNik/WhoFlips.

An AI scientist should not grade its own homework. Yet in the systems we examined, the agent that reviews the work usually comes from the same model family as the agent that produced it, or at least from the same vendor. Model evaluators are known to favour their own generations. Whether models trained alike also share blind spots is a conjecture, not a settled finding, but if they do, the reviewer inherits the author's. The record of what was flagged and what was waved through often sits in platform logs that nobody outside can replay. We present CrossAudit, a protocol for supervising autonomous research pipelines. It rests on three commitments. Each increment of work is audited by an agent from a different vendor against a rulebook a human wrote and versioned. Reports, verdicts, disputes and rulings are git commits, so the supervision history can be re-read and cited; raw model exchanges are not yet part of that record. Scripted checks run before any model does. Advisory judgement never gates the pipeline: a model blocks only by citing a rule, and no model may waive a deterministic failure. Blockers that survive a bounded number of revision rounds go to a person. We state the protocol as eight invariants. We describe a reference implementation built from GitHub Actions and a few hundred lines of Python, and report a live deployment of a closely related variant in a computational-chemistry pipeline. We also ran a seeded-defect trial (30 increments, 43 seeded defects, one run per configuration). A cross-vendor audit of our own repository then voided its blinding. We adopt that audit's findings and report the corrected results. The trial shows that two vendors read the same rulebook differently. It does not show that either is better. The strongest evidence here is the committed, uncontrolled record of cross-vendor audits of this paper itself.

The foundation of reproducible science lies in protocols that are precise, logically ordered, and executable. The autonomous generation of these protocols through natural language queries could greatly improve the efficiency of the reproduction process. However, current leading large language models (LLMs) often generate incomplete or inconsistent protocols, limiting their utility. To address this limitation, we first introduce SciRecipe, a large-scale dataset of over 12K structured protocols spanning 27 biological subfields and encompassing both comprehension and problem-solving tasks. To further improve protocol generation, we propose the "Sketch-and-Fill" paradigm, which separates analysis, structuring, and expression to ensure each step is explicit and verifiable. Complementing this, the structured component-based reward mechanism evaluates step granularity, action order, and semantic fidelity, aligning model optimization with experimental reliability. Building on these components, we develop Thoth, trained through a staged Knowledge-to-Action process that progresses from knowledge acquisition to operational reasoning and ultimately to robust, executable protocol generation. Across multiple benchmarks, Thoth consistently surpasses both proprietary and open-source LLMs, achieving significant improvements in step alignment, logical sequencing, and semantic accuracy. Our approach paves the way for reliable scientific assistants that bridge knowledge with experimental execution. All data, code, and models will be released publicly.

Retrieval-augmented generation (RAG) grounds answers in retrieved passages, yet relevance does not guarantee sufficiency: a topical passage may still fail to justify the answer. We study evidence sufficiency verification for selective RAG answering, in which a verifier receives a question, a candidate answer, and retrieved evidence and decides whether the evidence supports, refutes, or is insufficient for the answer, answering only when support is established. We present SURE-RAG, an aggregation protocol that treats evidence sufficiency as a set-level property: missing hops and unresolved conflicts cannot be detected by scoring passages independently. A shared claim-evidence verifier produces a local relation distribution for each (claim, passage) pair, which SURE-RAG aggregates into four interpretable answer-level feature blocks (coverage, relation strength, uncertainty, and retrieval), producing a three-way decision and an auditable selective score. We evaluate on HotpotQA-RAG v3, a controlled multi-hop benchmark, under an artifact-aware protocol (shortcut baselines, counterfactual swaps, no-oracle checks, and GPT-4o audits). Calibrated SURE-RAG attains 0.9075 Macro-F1 (raw 0.8951 +/- 0.0069), well above DeBERTa mean-pooling (0.6516) and a GPT-4o judge (0.7284), and on par with a strong concat cross-encoder (0.8888 +/- 0.0109) while remaining fully auditable. At 30% coverage, risk falls from 0.2588 to 0.1642, a 37% relative reduction. As a boundary-mapping experiment, we contrast SURE-RAG with GPT-4o on HaluBench unsafe detection: the ranking reverses (0.3343 vs. 0.7389 unsafe-F1), indicating that controlled sufficiency verification and natural hallucination detection are distinct problems.

Comprehensive and constructive evaluation protocols play an important role in the development of sophisticated text-to-video (T2V) generation models. Existing evaluation protocols primarily focus on temporal consistency and content continuity, yet largely ignore the dynamics of video content. Dynamics are an essential dimension for measuring the visual vividness and the honesty of video content to text prompts. In this study, we propose an effective evaluation protocol, termed DEVIL, which centers on the dynamics dimension to evaluate T2V models. For this purpose, we establish a new benchmark comprising text prompts that fully reflect multiple dynamics grades, and define a set of dynamics scores corresponding to various temporal granularities to comprehensively evaluate the dynamics of each generated video. Based on the new benchmark and the dynamics scores, we assess T2V models with the design of three metrics: dynamics range, dynamics controllability, and dynamics-based quality. Experiments show that DEVIL achieves a Pearson correlation exceeding 90% with human ratings, demonstrating its potential to advance T2V generation models. Code is available at https://github.com/MingXiangL/DEVIL.

Pass@k is widely used to report the reasoning performance of LLMs, but it often produces unstable and potentially misleading rankings, especially when the number of trials (samples) is limited and computational resources are constrained. We present a principled Bayesian evaluation framework that replaces Pass@k and average accuracy over N trials (avg@N) with posterior estimates of a model's underlying success probability and credible intervals, yielding stable rankings and a transparent decision rule for differences. Evaluation outcomes are modeled as categorical (not just 0/1) with a Dirichlet prior, giving closed-form expressions for the posterior mean and uncertainty of any weighted rubric and enabling the use of prior evidence when appropriate. Theoretically, under a uniform prior, the Bayesian posterior mean is order-equivalent to average accuracy (Pass@1), explaining its empirical robustness while adding principled uncertainty. Empirically, in simulations with known ground-truth success rates and on AIME'24/'25, HMMT'25, and BrUMO'25, the posterior-based procedure achieves faster convergence and greater rank stability than Pass@k and recent variants, enabling reliable comparisons at far smaller sample counts. The framework clarifies when observed gaps are statistically meaningful (non-overlapping credible intervals) versus noise, and it naturally extends to graded, rubric-based evaluations. Together, these results recommend replacing Pass@k for LLM evaluation and ranking with a posterior-based, compute-efficient protocol that unifies binary and non-binary evaluation while making uncertainty explicit. Source code is available at https://github.com/mohsenhariri/scorio

We introduce Sci-VBench, a comprehensive benchmark for evaluating knowledge- and reasoning-intensive video generation across scientific domains. It contains 1,253 expert-annotated examples spanning 60 subjects across four core disciplines: Natural Science, Healthcare, Humanities & Social Sciences, and Engineering. Each example requires models to generate temporally rich videos that demand scientific reasoning and knowledge-grounded synthesis, going beyond surface-level visual plausibility. We further establish a rubric-based evaluation protocol. Our analysis shows that, under this protocol, both non-expert human evaluators and MLLM-as-Judge systems can achieve relatively high agreement with expert judgments, supporting reproducible evaluation at scale. We benchmark 16 frontier proprietary and open-source models and find that, while automatic perceptual-quality scores cluster tightly across systems, performance on Prompt Grounding and Scientific and Causal Correctness varies substantially, with a pronounced proprietary-open-source gap. These findings show that advances in visual realism have not yet translated into reliable modeling of scientific and causal dynamics.

Multi-agent large language model (LLM) debate is often evaluated by whether final answers improve, but movement is not necessarily improvement: the same discussion can rescue an initially wrong majority or destroy an initially correct one. Standard final-accuracy evaluations conflate these opposing mechanisms. We introduce an auditable protocol for homogeneous debate on multiple-choice questions (MCQs) that records each run as a transition ledger over collapse, correction, onset, and signed intervention utility. On 6,925 MMLU-Pro debates, the protocol identifies 253 collapses and a parallel correction ledger that changes how interventions should be judged. Replay experiments reveal the central tradeoff: a leave-one-model-out probe-gated freeze prevents 29 collapses but loses 108 corrections under equal weights, so collapse prevention alone can recommend the wrong policy. A compact pre-debate 8-probe screen is a triage signal: its unadjusted family-level association with conditional-collapse risk is high (G=7, Spearman rho=0.893, exact two-sided p=0.0123), but initial-majority accuracy is a close comparator (rho=0.821; family partial rho=0.767, p=0.0877), so we do not treat it as calibrated or capability-adjusted prediction. Round-level traces localize many collapses to the first debate round, where early disagreement can precede both harmful cascades and useful recovery. We release replayable schemas, coders, audits, cost cards, and zero-API rebuild scripts so future model-scaffold rows can be compared under the same denominators and signed utility ledger.

As large language models (LLMs) become more powerful and are deployed more autonomously, it will be increasingly important to prevent them from causing harmful outcomes. Researchers have investigated a variety of safety techniques for this purpose, e.g. using models to review the outputs of other models, or red-teaming techniques to surface subtle failure modes. However, researchers have not evaluated whether such techniques still ensure safety if the model is itself intentionally trying to subvert them. In this paper, we develop and evaluate pipelines of safety techniques ("protocols") that are robust to intentional subversion. We investigate a scenario in which we want to solve a sequence of programming problems, using access to a powerful but untrusted model (in our case, GPT-4), access to a less powerful trusted model (in our case, GPT-3.5), and limited access to high-quality trusted labor. We investigate protocols that aim to never submit solutions containing backdoors, which we operationalize here as logical errors that are not caught by test cases. We investigate a range of protocols and test each against strategies that the untrusted model could use to subvert them. One protocol is what we call trusted editing. This protocol first asks GPT-4 to write code, and then asks GPT-3.5 to rate the suspiciousness of that code. If the code is below some suspiciousness threshold, it is submitted. Otherwise, GPT-3.5 edits the solution to remove parts that seem suspicious and then submits the edited code. Another protocol is untrusted monitoring. This protocol asks GPT-4 to write code, and then asks another instance of GPT-4 whether the code is backdoored, using various techniques to prevent the GPT-4 instances from colluding. These protocols improve substantially on simple baselines.

Deep research -- producing comprehensive, citation-grounded reports by searching and synthesizing information from hundreds of live web sources -- marks an important frontier for agentic systems. To rigorously evaluate this ability, four principles are essential: tasks should be (1) user-centric, reflecting realistic information needs, (2) dynamic, requiring up-to-date information beyond parametric knowledge, (3) unambiguous, ensuring consistent interpretation across users, and (4) multi-faceted and search-intensive, requiring search over numerous web sources and in-depth analysis. Existing benchmarks fall short of these principles, often focusing on narrow domains or posing ambiguous questions that hinder fair comparison. Guided by these principles, we introduce LiveResearchBench, a benchmark of 100 expert-curated tasks spanning daily life, enterprise, and academia, each requiring extensive, dynamic, real-time web search and synthesis. Built with over 1,500 hours of human labor, LiveResearchBench provides a rigorous basis for systematic evaluation. To evaluate citation-grounded long-form reports, we introduce DeepEval, a comprehensive suite covering both content- and report-level quality, including coverage, presentation, citation accuracy and association, consistency and depth of analysis. DeepEval integrates four complementary evaluation protocols, each designed to ensure stable assessment and high agreement with human judgments. Using LiveResearchBench and DeepEval, we conduct a comprehensive evaluation of 17 frontier deep research systems, including single-agent web search, single-agent deep research, and multi-agent systems. Our analysis reveals current strengths, recurring failure modes, and key system components needed to advance reliable, insightful deep research.

While Large Language Models (LLMs) have demonstrated proficiency in Deep Research or Wide Search, their capacity to solve highly complex questions-those requiring long-horizon planning, massive evidence gathering, and synthesis across heterogeneous sources-remains largely unexplored. We introduce Super Research, a task for complex autonomous research tasks that integrates (i) structured decomposition into a research plan, (ii) super wide retrieval for diverse perspectives, and (iii) super deep investigation to resolve uncertainties through iterative queries. To evaluate this capability, we curated a benchmark of 300 expert-written questions across diverse domains, each requiring up to 100+ retrieval steps and 1,000+ web pages to reconcile conflicting evidence. Super Research produces verifiable reports with fine-grained citations and intermediate artifacts (e.g., outlines and tables) to ensure traceable reasoning. Furthermore, we present a graph-anchored auditing protocol that evaluates Super Research along five dimensions: Coverage, Logical Consistency, Report Utility, Objectivity and Citation Health. While super-complex questions may be infrequent in standard applications, Super Research serves as a critical ceiling evaluation and stress test for LLM capabilities. A model's proficiency within Super Research acts as a powerful proxy for its general research competence; success here suggests the robustness necessary to navigate nearly any subordinate research task. Leaderboard is available at: https://cnsdqd-dyb.github.io/Super-Research-Benchmark/

Large language models (LLMs) are becoming increasingly capable mathematical collaborators, but static benchmarks are no longer sufficient for evaluating progress: they are often narrow in scope, quickly saturated, and rarely updated. This makes it hard to compare models reliably and track progress over time. Instead, we need evaluation platforms: continuously maintained systems that run, aggregate, and analyze evaluations across many benchmarks to give a comprehensive picture of model performance within a broad domain. In this work, we build on the original MathArena benchmark by substantially broadening its scope from final-answer olympiad problems to a continuously maintained evaluation platform for mathematical reasoning with LLMs. MathArena now covers a much wider range of tasks, including proof-based competitions, research-level arXiv problems, and formal proof generation in Lean. Additionally, we maintain a clear evaluation protocol for all models and regularly design new benchmarks as model capabilities improve to ensure that MathArena remains challenging. Notably, the strongest model, GPT-5.5, now reaches 98% on the 2026 USA Math Olympiad and 74% on research-level questions, showing that frontier models can now comfortably solve extremely challenging mathematical problems. This highlights the importance of continuously maintained evaluation platforms like MathArena to track the rapid progress of LLMs in mathematical reasoning.

Generating step-by-step "how-to" procedures is a key LLM capability: how-to advice is commonly requested in chatbots, and step-by-step planning is critical for reasoning over complex tasks. Yet, measuring and improving procedural validity at scale on real-world tasks remains challenging and understudied. To address this, we introduce How2Everything, a scalable framework to evaluate and improve goal-conditioned procedure generation. Our framework includes How2Mine, which mines 351K procedures from 980K web pages across 14 topics and readily scales to larger corpora. From this pool we build How2Bench, a 7K-example evaluation set balanced across topics. To reliably score model outputs, we develop How2Score, an evaluation protocol that uses an LLM judge to detect whether a generation contains any critical failure that would prevent achieving the goal. For low-cost, reproducible evaluation, we distill a frontier model into an open 8B model, achieving 80.5% agreement with human annotators. How2Bench reveals clear scaling trends across model sizes and training stages, providing signal early in pretraining. Finally, RL using How2Score as a reward improves performance on How2Bench by >10 points across three models without systematic regressions on standard benchmarks, with gains robust to superficial source-document memorization or format compliance. Taken together, How2Everything shows how pretraining web data can support a closed loop of capability evaluation and improvement at scale.

When evaluating large language models (LLMs) with multiple-choice question answering (MCQA), it is common to end the prompt with the string "Answer:" to facilitate automated answer extraction via next-token probabilities. However, there is no consensus on how to tokenize the space following the colon, often overlooked as a trivial choice. In this paper, we uncover accuracy differences of up to 11% due to this (seemingly irrelevant) tokenization variation as well as reshuffled model rankings, raising concerns about the reliability of LLM comparisons in prior work. Surprisingly, we are able to recommend one specific strategy -- tokenizing the space together with the answer letter -- as we observe consistent and statistically significant performance improvements. Additionally, it improves model calibration, enhancing the reliability of the model's confidence estimates. Our findings underscore the importance of careful evaluation design and highlight the need for standardized, transparent evaluation protocols to ensure reliable and comparable results.

As LLMs are deployed in high-stakes settings, users must judge the correctness of individual responses, often relying on model-generated justifications such as reasoning chains or explanations. Yet, no standard measure exists for whether these justifications help users distinguish correct answers from incorrect ones. We formalize this idea as error verifiability and propose v_{bal}, a balanced metric that measures whether justifications enable raters to accurately assess answer correctness, validated against human raters who show high agreement. We find that neither common approaches, such as post-training and model scaling, nor more targeted interventions recommended improve verifiability. We introduce two methods that succeed at improving verifiability: reflect-and-rephrase (RR) for mathematical reasoning and oracle-rephrase (OR) for factual QA, both of which improve verifiability by incorporating domain-appropriate external information. Together, our results establish error verifiability as a distinct dimension of response quality that does not emerge from accuracy improvements alone and requires dedicated, domain-aware methods to address.

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.

The evolution of Large Language Models (LLMs) towards autonomous agents has catalyzed progress in Deep Research. While retrieval capabilities are well-benchmarked, the post-retrieval synthesis stage--where agents must digest massive amounts of context and consolidate fragmented evidence into coherent, long-form reports--remains under-evaluated due to the subjectivity of open-ended writing. To bridge this gap, we introduce DeepSynth-Eval, a benchmark designed to objectively evaluate information consolidation capabilities. We leverage high-quality survey papers as gold standards, reverse-engineering research requests and constructing "Oracle Contexts" from their bibliographies to isolate synthesis from retrieval noise. We propose a fine-grained evaluation protocol using General Checklists (for factual coverage) and Constraint Checklists (for structural organization), transforming subjective judgment into verifiable metrics. Experiments across 96 tasks reveal that synthesizing information from hundreds of references remains a significant challenge. Our results demonstrate that agentic plan-and-write workflows significantly outperform single-turn generation, effectively reducing hallucinations and improving adherence to complex structural constraints.

As large language models (LLMs) become increasingly capable, it is prudent to assess whether safety measures remain effective even if LLMs intentionally try to bypass them. Previous work introduced control evaluations, an adversarial framework for testing deployment strategies of untrusted models (i.e., models which might be trying to bypass safety measures). While prior work treats a single failure as unacceptable, we perform control evaluations in a "distributed threat setting" -- a setting where no single action is catastrophic and no single action provides overwhelming evidence of misalignment. We approach this problem with a two-level deployment framework that uses an adaptive macro-protocol to choose between micro-protocols. Micro-protocols operate on a single task, using a less capable, but extensively tested (trusted) model to harness and monitor the untrusted model. Meanwhile, the macro-protocol maintains an adaptive credence on the untrusted model's alignment based on its past actions, using it to pick between safer and riskier micro-protocols. We evaluate our method in a code generation testbed where a red team attempts to generate subtly backdoored code with an LLM whose deployment is safeguarded by a blue team. We plot Pareto frontiers of safety (# of non-backdoored solutions) and usefulness (# of correct solutions). At a given level of usefulness, our adaptive deployment strategy reduces the number of backdoors by 80% compared to non-adaptive baselines.

Large language models can solve substantially harder reasoning problems with more inference-time compute. The term "test-time scaling," however, now covers diverse inference algorithms that extend deliberation along a single trajectory, sample completed candidates and aggregate them through voting or verification, or search over unfinished partial states. These algorithms differ in their statistical structure, compute accounting, and failure modes. Treating these procedures as interchangeable under a single scalar "budget," or reporting accuracy without the inference protocol that produced it, makes results difficult to compare across studies. We develop a systematic account of test-time scaling along three axes. First, we formalize test-time scaling as budgeted inference over the implicit prefix tree of an autoregressive model and distinguish three structural regimes: single-trajectory sequential scaling, leaf-level scaling with terminal reduction, and prefix-level scaling. Second, we treat the evaluated object as the entire inference system and develop evaluation principles that separate end-to-end system performance from candidate-bank diagnostics. We introduce an evaluation profile whose coordinates and simple functionals recover or bound common repeated-sampling metrics, and prescribe protocol-matched reporting of compute and uncertainty. Third, we specify reproducibility requirements for inference protocols, distinguishing exact replay from distributional reproducibility and identifying the artifacts needed to support each. We also organize the open-weight reasoning ecosystem by model-side and interface mechanisms, apply these principles to broad-knowledge, symbolic-reasoning, and competition-mathematics benchmarks, and assemble over 2 billion full reasoning traces for release with progressively richer verifier and token-level signals.

Given the widespread adoption and usage of Large Language Models (LLMs), it is crucial to have flexible and interpretable evaluations of their instruction-following ability. Preference judgments between model outputs have become the de facto evaluation standard, despite distilling complex, multi-faceted preferences into a single ranking. Furthermore, as human annotation is slow and costly, LLMs are increasingly used to make these judgments, at the expense of reliability and interpretability. In this work, we propose TICK (Targeted Instruct-evaluation with ChecKlists), a fully automated, interpretable evaluation protocol that structures evaluations with LLM-generated, instruction-specific checklists. We first show that, given an instruction, LLMs can reliably produce high-quality, tailored evaluation checklists that decompose the instruction into a series of YES/NO questions. Each question asks whether a candidate response meets a specific requirement of the instruction. We demonstrate that using TICK leads to a significant increase (46.4% to 52.2%) in the frequency of exact agreements between LLM judgements and human preferences, as compared to having an LLM directly score an output. We then show that STICK (Self-TICK) can be used to improve generation quality across multiple benchmarks via self-refinement and Best-of-N selection. STICK self-refinement on LiveBench reasoning tasks leads to an absolute gain of +7.8%, whilst Best-of-N selection with STICK attains +6.3% absolute improvement on the real-world instruction dataset, WildBench. In light of this, structured, multi-faceted self-improvement is shown to be a promising way to further advance LLM capabilities. Finally, by providing LLM-generated checklists to human evaluators tasked with directly scoring LLM responses to WildBench instructions, we notably increase inter-annotator agreement (0.194 to 0.256).

Computer-using agents (CUAs) enable task completion through natural interaction with operating systems and software interfaces. While script-based verifiers are widely adopted for evaluation, they suffer from limited scalability and inability to provide step-wise assessment. Reward models offer promising alternatives, but their effectiveness on CUA evaluation remains largely underexplored. To address this gap, we present CUARewardBench, comprising four key contributions: (1) First-ever Comprehensive CUA Reward Benchmark: We introduce the first benchmark for evaluating both outcome reward models (ORM) and process reward models (PRM) on CUA tasks, enabling systematic assessment across trajectory-level and step-level evaluation. (2) Diverse, Practical and Reliable Dataset: CUARewardBench encompasses trajectories from 10 software categories and 7 agent architectures with varying performance levels (25.9%-50.8% success rates). All trajectories are expertly annotated through carefully designed protocols, with rigorous quality control to ensure reliability and practical applicability. (3) Comprehensive Analysis and Insights: Through extensive experiments across 7 vision-language models and 3 prompt templates, we reveal critical limitations of current CUA RMs, including insufficient visual reasoning capabilities, knowledge deficiencies, and the superiority of general VLMs over specialized CUA models for reward evaluation. (4) Unanimous Prompt Ensemble (UPE): Based on the insights from our comprehensive analysis, we propose UPE, a novel ensemble method that significantly enhances reward model reliability through strict unanimous voting and strategic prompt-template configurations. UPE achieves 89.8% precision and 93.3% NPV for ORM, and 81.7% precision and 85.1% NPV for PRM, substantially outperforming single VLMs and traditional ensemble approaches.

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.

Red-team evaluations of AI models support some claims and not others, and the boundary between the two is calculable rather than merely a matter of judgment. We define the evidential ceiling of an evaluation as the largest factor by which one result can move belief under a fixed testing budget, derive it in closed form for the benchmark null result, and use it to locate that boundary exactly. We find that above a calculable harm rate, a benchmark of modest size certifies a category to a stated evidentiary standard, and a clean sheet is then the stronger of the two possible observations, outweighing a single reproduced failure. Below that rate, no passive benchmark of feasible size provides the specified evidence of safety under the fixed scoring rule and approximately independent trial structure. The crossing between the two regimes has a closed form. The bound is not specific to benchmarks: written in terms of a procedure's hypothesis conditioned elicitation rates, it covers adaptive and automated red teaming as well, and shows that discrimination between the hypotheses rather than attack success is what determines evidential worth. Auditing eight evaluation suites against the boundary, we find that current benchmarks are adequate for high-frequency harm categories and several orders of magnitude short for rare, catastrophic ones. Safety benchmarks are not uninformative. They are informative about a specific and computable set of propositions, and the discipline they need is to state which.

Automated research systems support literature synthesis, ideation, experiments, writing, and peer review, but their evaluation is dispersed across tasks, benchmarks, and studies that are difficult to compare directly. We review this literature from the perspective of evaluation design and evidence, covering six targets: literature synthesis, research ideation, executable workflows, scholarly writing and communication, automatic peer review, and end-to-end research. We compare task construction, evidence sources, evaluators, and scoring procedures to explain the capabilities assessed by different designs. Our synthesis highlights three recurring lessons: output checks, process checks, and human studies provide complementary information; evaluator calibration is specific to the property being assessed; and resource budgets and attempt selection are integral to interpreting performance comparisons. We identify diagnostic evaluation designs and documented gaps in supporting evidence, and translate these comparisons into reporting and audit recommendations for specific evaluation settings. The survey helps readers navigate existing evaluations, select appropriate benchmarks, and design subsequent studies.

The main issue with most evaluation schemes today is their "static" nature: the same problems are reused repeatedly, allowing for memorization, format exploitation, and eventual saturation. To measure genuine AI progress, we need evaluation that is robust by construction, not by post-hoc detection. In response, we propose VeRA (Verified Reasoning Data Augmentation), a framework that converts benchmark problems into executable specifications, comprising (i) a natural language template with placeholder slots, (ii) a coherent generator that samples valid configurations, and (iii) a deterministic verifier that validates parameters and calculates the corresponding correct answers for each configuration. From a single seed problem, VeRA automatically creates unlimited verified variants with reliable labels at near-zero marginal cost without human involvement. VeRA operates in two complementary modes. VeRA-E (equivalent) rewrites problems while keeping the underlying logic intact, useful for detecting memorization versus genuine reasoning. VeRA-H (hardened) systematically increases complexity while remaining verifiable, enabling reliable creation and labelling of fresh difficult tasks at the boundary of intelligence. Evaluating 16 frontier models with VeRA, we find: (i) VeRA-E improves evaluation quality and reveals contamination patterns. (ii) VeRA-H enables human-free generation of hard tasks with reliable labels. (iii) VeRA establishes verified benchmarks as a general paradigm. VeRA reconceptualizes benchmarks from static objects used until exhausted, to executable specifications generating fresh, verified instances on demand, enhancing robustness and cost-effectiveness for evaluation. With VeRA, we envision that evaluation in any verifiable domain can scale indefinitely without sacrificing label integrity. To stimulate future research, we have open-sourced all code and datasets.

Deep Research Agents (DRAs) have demonstrated remarkable capabilities in autonomous information retrieval and report generation, showing great potential to assist humans in complex research tasks. Current evaluation frameworks primarily rely on LLM-generated references or LLM-derived evaluation dimensions. While these approaches offer scalability, they often lack the reliability of expert-verified content and struggle to provide objective, fine-grained assessments of critical dimensions. To bridge this gap, we introduce Wiki Live Challenge (WLC), a live benchmark that leverages the newest Wikipedia Good Articles (GAs) as expert-level references. Wikipedia's strict standards for neutrality, comprehensiveness, and verifiability serve as a great challenge for DRAs, with GAs representing the pinnacle of which. We curate a dataset of 100 recent Good Articles and propose Wiki Eval, a comprehensive evaluation framework comprising a fine-grained evaluation method with 39 criteria for writing quality and rigorous metrics for factual verifiability. Extensive experiments on various DRA systems demonstrate a significant gap between current DRAs and human expert-level Wikipedia articles, validating the effectiveness of WLC in advancing agent research. We release our benchmark at https://github.com/WangShao2000/Wiki_Live_Challenge

Deep research agents increasingly automate complex information-seeking tasks, producing evidence-grounded reports via multi-step reasoning, tool use, and synthesis. Their growing role demands scalable, reliable evaluation, positioning LLM-as-judge as a supervision paradigm for assessing factual accuracy, evidence use, and reasoning quality. Yet the reliability of these judges for deep research agents remains poorly understood, posing a critical meta-evaluation problem: before deploying LLM judges to supervise research agents, we must first evaluate the judges themselves. Existing meta-evaluations fall short in two ways: (1) reliance on coarse, subjective human-preference agreement; (2) focus on instruction-following or verifiable tasks, leaving open-ended agent executions unexplored. To address these gaps, we introduce REFLECT (REliable Fine-grained LLM judge Evaluation via Controlled inTervention), a meta-evaluation benchmark targeting fine-grained failure detection in agentic environments. REFLECT defines a detailed taxonomy of process- and outcome-level failure modes, instantiated by performing controlled and localized interventions on quality-screened agent execution traces. This yields verifiable, comprehensive, and fine-grained instances for validating the judge models. Our experiments show that current LLM judges remain unreliable: even the best-performing models achieve overall accuracies below 55% across reasoning, tool-use, and report-quality failures, with especially poor performance on evidence verification. Together, our taxonomy and findings expose systematic judge limitations, reveal tradeoffs in cost and reliability, and offer actionable guidance for building more reliable evaluation pipelines for deep research agents.

Large Language Models (LLMs) are increasingly used for clinical decision support, where hallucinations and unsafe suggestions may pose direct risks to patient safety. These risks are particularly challenging as they often manifest as subtle clinical errors that evade detection by generic metrics, while expert-authored fine-grained rubrics remain costly to construct and difficult to scale. In this paper, we propose a retrieval-augmented multi-agent framework designed to automate the generation of instance-specific evaluation rubrics. Our approach grounds evaluation in authoritative medical evidence by decomposing retrieved content into atomic facts and synthesizing them with user interaction constraints to form verifiable, fine-grained evaluation criteria. Evaluated on HealthBench, our framework achieves a Clinical Intent Alignment (CIA) score of 60.12%, a statistically significant improvement over the GPT-4o baseline (55.16%). In discriminative tests, our rubrics yield a mean score delta (μ_Δ = 8.658) and an AUROC of 0.977, nearly doubling the quality separation achieved by GPT-4o baseline (4.972). Beyond evaluation, our rubrics effectively guide response refinement, improving quality by 9.2% (from 59.0% to 68.2%). This provides a scalable and transparent foundation for both evaluating and improving medical LLMs. The code is available at https://anonymous.4open.science/r/Automated-Rubric-Generation-AF3C/.

The evaluation of generative models for natural image tasks has been extensively studied. Similar protocols and metrics are used in cases with unique particularities, such as Handwriting Generation, even if they might not be completely appropriate. In this work, we introduce three measures tailored for HTG evaluation, HTG_{HTR} , HTG_{style} , and HTG_{OOV} , and argue that they are more expedient to evaluate the quality of generated handwritten images. The metrics rely on the recognition error/accuracy of Handwriting Text Recognition and Writer Identification models and emphasize writing style, textual content, and diversity as the main aspects that adhere to the content of handwritten images. We conduct comprehensive experiments on the IAM handwriting database, showcasing that widely used metrics such as FID fail to properly quantify the diversity and the practical utility of generated handwriting samples. Our findings show that our metrics are richer in information and underscore the necessity of standardized evaluation protocols in HTG. The proposed metrics provide a more robust and informative protocol for assessing HTG quality, contributing to improved performance in HTR. Code for the evaluation protocol is available at: https://github.com/koninik/HTG_evaluation.

Reasoning protocols such as Chain of Thought (CoT) and Tree of Thought (ToT) organize internal deliberation but lack an explicit mechanism for external questioning that elicits self-revision. We present FOR-Prompting (From Objection to Revision Prompting), an asymmetric protocol where a Defender proposes an answer, an Objectioner raises question-style objections with no direct fixes, and a Host enforces consistency and closure. On GSM8K we observe about a 22% point gain over single-prompt and accuracy on par with CoT, with more than 10% higher ratings in reasoning and coherence from a uniform GPT 4.1 judge. FOR-Prompting also corrects mistakes without tools or human supervision on tricky queries, and improves performance for small-scale model (approx. 19% accuracy improved on Llama3.2:1b for GSM8K task), highlighting promise for small models and on personal device use. Beyond factual QA, qualitative analyses on open-ended tasks show enhanced exploration and refinement, with dialogue traces that make assumptions and trade-offs explicit. The protocol is model agnostic and operates purely at the prompt level through role-structured turns, so it works with hosted and local models of different sizes without retraining, and it supports large-scale study of objection-guided reasoning.