new

Get trending papers in your email inbox!

Subscribe

Daily Papers

byAK and the research community

Nov 17

Reinforcement Learning for LLM-based Multi-Agent Systems through Orchestration Traces

As large language model (LLM) agents evolve from isolated tool users into coordinated teams, reinforcement learning (RL) must optimize not only individual actions but also how work is spawned, delegated, communicated, aggregated, and stopped. This paper studies RL for LLM-based multi-agent systems through orchestration traces: temporal interaction graphs whose events include sub-agent spawning, delegation, communication, tool use, return, aggregation, and stopping decisions. Using this lens, we identify three technical axes. First, reward design spans eight families, including orchestration rewards for parallelism speedup, split correctness, and aggregation quality. Second, reward and credit signals attach to eight credit- or signal-bearing units from token to team; explicit counterfactual message-level credit remains especially sparse in our curated pool. Third, orchestration learning decomposes into five sub-decisions: when to spawn, whom to delegate to, how to communicate, how to aggregate, and when to stop. In our curated pool as of May 4, 2026, we found no explicit RL training method for the stopping decision. We connect academic methods to public industrial evidence from Kimi Agent Swarm, OpenAI Codex, and Anthropic Claude Code. The resulting scale gap is a gap between publicly reported deployment envelopes and open academic evaluation regimes, not independent verification of industrial training traces. We release the artifact at https://github.com/xxzcc/awesome-llm-mas-rl, including an 84-entry tagged paper pool, a 32-record exclusion log, scripted corpus statistics, and a minimal JSON schema for replayable orchestration traces.

  • 1 authors
·
May 3 3

TacMind: Benchmarking and Modeling Tactile Interaction States for Robotic Manipulation

Tactile sensing provides robots with direct observations of contact physics, local geometry, and interaction dynamics, making it a vital source of feedback for contact-rich manipulation. However, existing tactile manipulation studies are often developed and evaluated within specific task settings, making it difficult to systematically compare how tactile information contributes under different interaction conditions. In this work, we develop a benchmark in simulation comprising 128 tasks organized around three complementary evaluation regimes: Precise, Recognition, and Dynamic. The benchmark further includes controlled static and dynamic task variants and 12 perturbation settings for evaluating robustness under changes in visual and physical conditions. These settings motivate studying tactile information as a temporally evolving interaction representation rather than only as an instantaneous observation. We therefore propose an action-conditioned latent predictive world model that integrates tactile history and learns a tactile state through prediction of future tactile latents. Building on this representation, we design a model that conditions action generation on the learned tactile state and uses predicted future tactile information to refine candidate actions through predictive tactile reflection. Experiments on the proposed benchmark show consistent gains from incorporating tactile information across the evaluated regimes. Project Page: https://tacmind-bench.github.io/

  • 6 authors
·
Oct 29

'Si'multaneous 'S'patial-'T'emporal Message Passing for Dynamic Graph Representation Learning

Dynamic graph neural networks (DGNNs) that operate on snapshot sequences typically fall into one of two categories. Temporal-first approaches build per-node temporal embeddings and only afterwards perform spatial aggregation, whereas Spatial-first approaches invert this order, feeding the output of a graph convolution into a downstream temporal module. In either case, the rigid sequencing forces the second stage to consume an already-compressed summary produced by the first, ruling out joint reasoning over topology and evolution; concretely, the message-passing operator never gets to weight a neighbor's contribution by that neighbor's past trajectory. This paper introduces SiST-GNN (Simultaneous Spatial-Temporal GNN), which fuses the two signals inside a single message-passing operation rather than chaining them. Concretely, at each snapshot we maintain a recurrent hidden state per node that summarises its history, pair it with the node's current feature vector, and treat the pair as two nodes joined by a cross-time edge; running a standard graph convolution on this temporally augmented graph yields the updated representation. Our empirical study spans nine public baselines and fourteen model-dataset combinations, covering both fixed-split and live-update evaluation regimes. Across every public benchmark, SiST-GNN sets a new state of the art in link prediction task over the strongest prior method by 109--277% in the fixed-split setting and by 68--194% in the live-update setting. We additionally construct three dynamic node-classification tasks by discretising the underlying continuous-time event streams; here SiST-GNN beats the leading discrete-time (DTDG) baseline by 7--22% and matches continuous-time (CTDG) methods that consume the raw events directly.

  • 2 authors
·
May 24

Test-Time Scaling in Reasoning LLMs: Inference Regimes, Evaluation, and Reproducibility

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.

  • 14 authors
·
Aug 3

PeerPrism: Peer Evaluation Expertise vs Review-writing AI

Large Language Models (LLMs) are increasingly used in scientific peer review, assisting with drafting, rewriting, expansion, and refinement. However, existing peer-review LLM detection methods largely treat authorship as a binary problem-human vs. AI-without accounting for the hybrid nature of modern review workflows. In practice, evaluative ideas and surface realization may originate from different sources, creating a spectrum of human-AI collaboration. In this work, we introduce PeerPrism, a large-scale benchmark of 20,690 peer reviews explicitly designed to disentangle idea provenance from text provenance. We construct controlled generation regimes spanning fully human, fully synthetic, and multiple hybrid transformations. This design enables systematic evaluation of whether detectors identify the origin of the surface text or the origin of the evaluative reasoning. We benchmark state-of-the-art LLM text detection methods on PeerPrism. While several methods achieve high accuracy on the standard binary task (human vs. fully synthetic), their predictions diverge sharply under hybrid regimes. In particular, when ideas originate from humans but the surface text is AI-generated, detectors frequently disagree and produce contradictory classifications. Accompanied by stylometric and semantic analyses, our results show that current detection methods conflate surface realization with intellectual contribution. Overall, we demonstrate that LLM detection in peer review cannot be reduced to a binary attribution problem. Instead, authorship must be modeled as a multidimensional construct spanning semantic reasoning and stylistic realization. PeerPrism is the first benchmark evaluating human-AI collaboration in these settings. We release all code, data, prompts, and evaluation scripts to facilitate reproducible research at https://github.com/Reviewerly-Inc/PeerPrism.

  • 6 authors
·
Apr 15

StereoGenBench: A Synthetic Multi-Camera Benchmark for Stereo Generation under Controlled Baseline Regimes

Stereo image and video generation, stereo geometry estimation, and condition-controlled view synthesis require paired data in which the variables that determine binocular geometry -- camera baseline, intrinsics, scene depth, and camera motion -- are known and controllable. Existing stereo resources provide subsets of these variables, but resources commonly used for stereo generation evaluation do not, to our knowledge, provide scene-paired, calibrated multi-baseline right-view ground truth with jointly recorded intrinsics, dense metric depth, and per-frame poses in a single controlled source. We introduce StereoGenBench, a synthetic Unreal Engine benchmark designed to make baseline-regime sensitivity and target-camera consistency measurable under matched scene content. Each scene is rendered with a rigid six-camera lateral array, yielding up to 15 calibrated view pairs; adjacent baselines are sampled from inter-pupillary to wide-baseline regimes; focal length is sampled independently; and every view is released with RGB, metric depth, intrinsics, per-pair baselines, and per-frame poses. The splits include two evaluation families for narrow and wide baseline regimes and a train-only family for broader all-pairs coverage. We release the dataset, evaluation code, reference results, Croissant metadata, and generation code/configuration for extension with compatible assets. The dataset is available at https://huggingface.co/datasets/stereo-dataset/stereo-dataset

  • 3 authors
·
May 21

Evaluation of Multi-Turn Consistency in LLM Agents: Survival Analysis and Failure-Rationale Taxonomy

Large language model (LLM) agents may perform well on isolated tasks yet drift into inconsistency over extended interaction. We evaluate temporal consistency in a controlled 20-step multi-agent setting inspired by delayed-gratification studies. At each step, an agent chooses between continuing to delay a reward or claiming it immediately (terminating the episode). Across a full-factorial manipulation of social visibility (private vs public), persona stressors, and deliberation policy, we run 84,540 trajectories spanning 8 model families. Treating the first reward-claim as a time-to-event outcome, we estimate Kaplan-Meier survival curves and fit discrete-time hazard regression to quantify how experimental factors shift failure risk over time. Then, to analyze rationales and language patterns associated with failure, we build a seven-category taxonomy from 13,780 deliberation traces from agents who choose to terminate the episode, using an LLM-assisted labeling paired with human audit (κ=0.83). Rationale profiles change systematically with time and context: early failures are more impulse-driven, later failures more fatigue- and cost-benefit-framed, while public settings increase norm-oriented justifications. We also find a deliberation-inconsistency association: among failures, longer deliberation correlates with higher rates of intra-rationale contradiction (simultaneous pro-delay and pro-claim statements), challenging the assumption that more reasoning text implies greater consistency. Together, the survival and rationale analyses reveal distinct temporal reliability regimes and model-specific "failure fingerprints", offering an evaluation lens for diagnosing inconsistency in multi-turn agent behavior.

  • 3 authors
·
Aug 25

IND-Eval: Aligning Evaluation and Simulation for Industrial Time Series Foundation Models

We introduce IND-Eval, a public benchmark that organizes industrial time series across domains into source-specific forecasting tasks. Relative to established benchmarks such as GIFT-Eval, it emphasizes finer sampling resolutions and a greater prevalence of multivariate forecasting. Evaluating Chronos-2, TiRex-2, and TimesFM-3, we find that TimesFM-3 achieves the lowest source-balanced aggregate nMASE and nCRPS and leads seven of nine sources under each metric; source-specific exceptions reinforce the need for domain-representative evaluation for model selection. These findings also motivate improving domain alignment during pretraining, where real industrial data are often too scarce to close domain-specific gaps directly; we therefore investigate whether a domain's statistical structure can guide simulator design. In a case study on two offshore well-monitoring regimes, we improve a general simulator's coverage using an oscillator-augmented simulation pipeline. This lowers forecasting error in the target domain but reduces performance on the subset of GIFT-Eval with natively univariate targets. A 50/50 mixture of the general and augmented simulation recipes achieves the lowest error among the three training recipes for all four event-metric pairs in the case study while recovering most of the lost GIFT-Eval performance. Together, these results illustrate how domain-representative evaluation can guide model selection and the composition of simulated data while revealing a trade-off between specialization and generality.

cognitedata cognitedata
·
Oct 11

Critical Evaluation of Quantum Machine Learning for Adversarial Robustness

Quantum Machine Learning (QML) integrates quantum computational principles into learning algorithms, offering improved representational capacity and computational efficiency. Nevertheless, the security and robustness of QML systems remain underexplored, especially under adversarial conditions. In this paper, we present a systematization of adversarial robustness in QML, integrating conceptual organization with empirical evaluation across three threat models-black-box, gray-box, and white-box. We implement representative attacks in each category, including label-flipping for black-box, QUID encoder-level data poisoning for gray-box, and FGSM and PGD for white-box, using Quantum Neural Networks (QNNs) trained on two datasets from distinct domains: MNIST from computer vision and AZ-Class from Android malware, across multiple circuit depths (2, 5, 10, and 50 layers) and two encoding schemes (angle and amplitude). Our evaluation shows that amplitude encoding yields the highest clean accuracy (93% on MNIST and 67% on AZ-Class) in deep, noiseless circuits; however, it degrades sharply under adversarial perturbations and depolarization noise (p=0.01), dropping accuracy below 5%. In contrast, angle encoding, while offering lower representational capacity, remains more stable in shallow, noisy regimes, revealing a trade-off between capacity and robustness. Moreover, the QUID attack attains higher attack success rates, though quantum noise channels disrupt the Hilbert-space correlations it exploits, weakening its impact in image domains. This suggests that noise can act as a natural defense mechanism in Noisy Intermediate-Scale Quantum (NISQ) systems. Overall, our findings guide the development of secure and resilient QML architectures for practical deployment. These insights underscore the importance of designing threat-aware models that remain reliable under real-world noise in NISQ settings.

  • 5 authors
·
Nov 24, 2025

SimplexUQ: An Evaluation Framework and Benchmark for Conformal Uncertainty on Simplex-Valued Predictions

Conformal prediction guarantees marginal coverage, but a single calibration threshold can still spread that coverage unevenly, over-covering easy regions and under-covering hard ones. SimplexUQ is, to our knowledge, the first benchmark and reproducible protocol for measuring this allocation problem on simplex-valued predictions; it compares existing conformal wrappers rather than proposing a new one. Its task suite, SimplexTasks-12, combines six controlled synthetic regimes with six frozen-predictor real tasks spanning class probabilities, topic mixtures, spectral abundances, cell-type fractions, age distributions, and emotion mixtures. Each comparison fixes the predictor, score, and response-free stratification map, varies only the wrapper, and reports marginal coverage, worst-stratum coverage, max disparity, and within-task radius and compute. Global calibration can look valid while failing badly: on CIFAR-10 it attains 0.900 marginal coverage but only 0.542 in the worst entropy stratum, and Mondrian calibration raises that stratum to 0.886 while reducing max disparity from 0.358 to 0.022. No wrapper dominates, however. Under smooth synthetic heterogeneity, several repairs are competitive; fixed-map analyses show that rankings depend on the evaluation groups and protocol; and in a 12-task comparison, Mondrian has lower disparity on its single target partition for all 12 tasks, whereas BatchMVP has lower disparity over overlapping groups on five. These are empirical comparisons, not new coverage guarantees. A controlled predictor-bias sweep shows that removing predictor bias only partly reduces global-threshold disparity. We release task cards, result provenance, permitted derived arrays, and rebuild instructions, and treat wrapper selection as a diagnostic comparison rather than a universal ranking.

  • 3 authors
·
Sep 29

nnActive: A Framework for Evaluation of Active Learning in 3D Biomedical Segmentation

Semantic segmentation is crucial for various biomedical applications, yet its reliance on large annotated datasets presents a bottleneck due to the high cost and specialized expertise required for manual labeling. Active Learning (AL) aims to mitigate this challenge by querying only the most informative samples, thereby reducing annotation effort. However, in the domain of 3D biomedical imaging, there is no consensus on whether AL consistently outperforms Random sampling. Four evaluation pitfalls hinder the current methodological assessment. These are (1) restriction to too few datasets and annotation budgets, (2) using 2D models on 3D images without partial annotations, (3) Random baseline not being adapted to the task, and (4) measuring annotation cost only in voxels. In this work, we introduce nnActive, an open-source AL framework that overcomes these pitfalls by (1) means of a large scale study spanning four biomedical imaging datasets and three label regimes, (2) extending nnU-Net by using partial annotations for training with 3D patch-based query selection, (3) proposing Foreground Aware Random sampling strategies tackling the foreground-background class imbalance of medical images and (4) propose the foreground efficiency metric, which captures the low annotation cost of background-regions. We reveal the following findings: (A) while all AL methods outperform standard Random sampling, none reliably surpasses an improved Foreground Aware Random sampling; (B) benefits of AL depend on task specific parameters; (C) Predictive Entropy is overall the best performing AL method, but likely requires the most annotation effort; (D) AL performance can be improved with more compute intensive design choices. As a holistic, open-source framework, nnActive can serve as a catalyst for research and application of AL in 3D biomedical imaging. Code is at: https://github.com/MIC-DKFZ/nnActive

  • 9 authors
·
Nov 24, 2025

TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders

Tabular encoders are usually evaluated inside task-specific end-to-end pipelines, so models from different training paradigms are difficult to compare directly even when they operate on similar tabular signals. We introduce TRL-Bench, a multi-granular tabular representation learning (TRL) benchmark that standardizes cross-paradigm representation-level evaluation: each encoder exports row-, column-, or table embeddings through its supported wrapper, and shared lightweight heads probe them across three suites: TRL-CTbench (column/table), TRL-Rbench (row), and TRL-DLTE (compositional Data-Lake Table Enrichment spanning all three granularities). To support this standardized setting, we release curated benchmark assets and task reformulations, including 50 OpenML tables with 123 verified targets, 16 row-pair linkage rewrites, and a 47,772-table DLTE lake derived from 1,379 parent tables. Across 20 models and 16 tasks, TRL-Bench shows that once downstream conditions are standardized, encoder quality is capability-specific rather than captured by a single leaderboard. In TRL-CTbench, generic text encoders often lead on tasks with strong surface-text signal, while tabular specialists win where their pretraining objective aligns with the task. In TRL-Rbench, within-table prediction and cross-table linkage favor different training regimes, with atomic linkage performance correlating strongly with the row-matching stage of DLTE pipelines. In TRL-DLTE, the strongest pipelines combine capability-matched specialists rather than reuse a single encoder, and top end-to-end quality depends on non-additive compositional fit rather than per-stage marginal rank alone. TRL-Bench provides a common protocol for measuring reusable signal in exported tabular representations under shared downstream conditions. Code and data: https://github.com/LOGO-CUHKSZ/TRL-Bench

Regimes: An Auditable, Held-Out-Gated Improvement Loop Demonstrated on LongMemEval with ActiveGraph

Autonomous improvement loops are hard to trust because the improvement process is usually external scaffolding bolted onto the agent: failures go unlogged, diagnoses cannot be replayed, and promote-or-discard decisions land in a side database rather than the agent's own history. We show that an event-sourced agent runtime removes that friction and turns controlled improvement into a first-class workflow. When the agent's state is a deterministic projection of an append-only event log, failures are recorded, a run replays exactly from its log, candidate patches scope to typed pipeline seams, gates are auditable, and every promotion or discard is itself an event. We demonstrate this with Regimes, a loop on the ActiveGraph runtime that diagnoses failed evaluations, proposes a repair at a pipeline point, and promotes it only after static checks, sandbox execution, in-sample evaluation, and held-out validation. The loop is target-agnostic: the same control flow runs against different tasks through a common interface. On LongMemEval-S the dominant failure is not retrieval but reconciliation: the evidence is already in the assembled context, yet the reader answers incorrectly. Across five seeded held-out splits, Regimes discovers reader-prompt repairs that improve final held-out accuracy by +0.05 to +0.10 in four splits and +0.01 in one over-promotion split; two splits are individually significant (seed 5 unadjusted for its sequential promotion structure), and the pooled count is descriptive only, since the splits share one 500-question pool. The durable contributions are ActiveGraph as an auditable substrate that makes controlled improvement loops tractable, the held-out-gated loop it supports, the failure-regime taxonomy routing each failure to a pipeline location (whose marginal value over an unrouted baseline is the primary open question), and the prompt-as-discovery-probe hypothesis.

  • 1 authors
·
Jun 7

Going with the Speed of Sound: Pushing Neural Surrogates into Highly-turbulent Transonic Regimes

The widespread use of neural surrogates in automotive aerodynamics, enabled by datasets such as DrivAerML and DrivAerNet++, has primarily focused on bluff-body flows with large wakes. Extending these methods to aerospace, particularly in the transonic regime, remains challenging due to the high level of non-linearity of compressible flows and 3D effects such as wingtip vortices. Existing aerospace datasets predominantly focus on 2D airfoils, neglecting these critical 3D phenomena. To address this gap, we present a new dataset of CFD simulations for 3D wings in the transonic regime. The dataset comprises volumetric and surface-level fields for around 30,000 samples with unique geometry and inflow conditions. This allows computation of lift and drag coefficients, providing a foundation for data-driven aerodynamic optimization of the drag-lift Pareto front. We evaluate several state-of-the-art neural surrogates on our dataset, including Transolver and AB-UPT, focusing on their out-of-distribution (OOD) generalization over geometry and inflow variations. AB-UPT demonstrates strong performance for transonic flowfields and reproduces physically consistent drag-lift Pareto fronts even for unseen wing configurations. Our results demonstrate that AB-UPT can approximate drag-lift Pareto fronts for unseen geometries, highlighting its potential as an efficient and effective tool for rapid aerodynamic design exploration. To facilitate future research, we open-source our dataset at https://huggingface.co/datasets/EmmiAI/Emmi-Wing.

  • 8 authors
·
Nov 26, 2025

Benchmarking NACTI Species Recognition in Long-Tailed Regimes

As with most ``in the wild'' collections of the natural world, the North America Camera Trap Images (NACTI) dataset exhibits long-tailed class imbalance, with the largest class covering over 50% of its 3.7M images. Building on the PyTorch Wildlife model, we systematically evaluate Long-Tail Recognition (LTR) methodologies to benchmark species recognition performance, including specialised loss functions and LTR-sensitive regularisation. Our optimised configuration achieves state-of-the-art 99.40% Top-1 accuracy on the NACTI test split, significantly outperforming standard baselines and previously reported top performances. To assess robustness under domain shifts (e.g., night-time captures, occlusion, motion-blur), we extend our evaluation across three independent reduced-bias test sets (including ENA-Detection, Caltech Camera Traps and Missouri Camera Traps). Across these out-of-distribution (OOD) evaluations, our LTR-enhanced model consistently demonstrates substantially stronger generalisation capabilities compared to standard cross-entropy approaches. However, qualitative and quantitative analyses underline that current LTR optimisations cannot fully overcome representational bottlenecks, resulting in catastrophic predictive breakdown for rare `Tail' classes under severe domain shift. For maximum reproducibility, all dataset splits, key code, and network weights are published with this paper at https://github.com/ZehuaLiuY/Species-Classification.

  • 2 authors
·
Jul 19

Harness the Memory: A Holistic Evaluation of Memory Substrates in Memory Agents

Memory is becoming core infrastructure for long-horizon LLM agents, yet existing evaluations offer limited guidance on which memory substrate, namely the underlying medium in which memory is represented and stored, should be used under different operating regimes. We present a controlled harness evaluation of memory substrates for memory-augmented agents, covering dense and sparse indices, text records, structural stores, hierarchical stores, refinement-based memories, parametric updates, and activation-compatible context mechanisms. Across three backbone models and four benchmark suites spanning user-centric question answering and agent-centric decision-making, we instrument 26 performance and efficiency metrics under a unified harness. Our results show that no single substrate consistently dominates: broad retrieval benefits long-context factual QA, while excessive retrieval can harm sequential decision-making by shifting attention away from action-critical context. Scalability introduces a further routing axis, as substrates that perform well at moderate history lengths can become costly or brittle at longer horizons. These findings motivate substrate routing as a necessary component of adaptive agent memory systems and provide empirical guidance for designing efficient, reliable, and regime-aware long-term memory for LLM agents. Code will be made available upon acceptance.

  • 15 authors
·
Aug 14 2

CreativityPrism: A Cross-Domain Evaluation Framework for Large Language Model Creativity

Creativity is often seen as a hallmark of human intelligence. While large language models(LLMs) are increasingly perceived as generating creative text, there is still no cross-domain and scalable framework to evaluate their creativity across diverse scenarios. Existing methods of LLM creativity evaluation either heavily rely on humans, limiting speed and scalability, or are fragmented across different domains and different definitions of creativity. To address this gap, we propose CreativityPrism, an evaluation and analysis framework that consolidates eight tasks from three domains: divergent thinking, creative writing, and logical reasoning, into a taxonomy of creativity that emphasizes three dimensions: quality, novelty, and diversity of LLM generations. The framework is designed to be scalable with reliable automatic evaluation judges that have been validated against human annotations. We evaluate 17 state-of-the-art (SoTA) LLMs on CreativityPrism and find that while frontier-scale LLMs dominate creative writing and logical reasoning tasks by a .10 (or 15%) lead over locally-deployable open models, they offer no significant advantage in divergent thinking, a domain much less explored in existing post-training regimes. Our analysis also shows that high performance in one creative dimension or domain rarely generalizes to others; specifically, novelty metrics often show weak or negative correlations with other metrics. This fragmentation confirms that a cross-domain, multi-dimensional framework like CreativityPrism is essential for any meaningful assessment of LLM creativity.

  • 12 authors
·
Jun 30

ARB: A Matched Authorship-Rewriting Benchmark Dataset for AI-Text Detector Evaluation

Standard AI-text detection benchmarks compare human-written text against text generated directly by large language models (LLMs). While prior work has shown that rewriting and paraphrasing can degrade detector performance, it remains unclear whether performance measured on this conventional benchmark predicts detector behavior when human-authored content is rewritten by an LLM. To address this gap, we introduce Authorship-Rewriting Benchmark (ARB), built from 1,800 human source texts (600 each from XSum, WritingPrompts, and OpenWebText) and four open-weight generators (Llama-3.2-3B, Qwen2.5-7B, Mistral-7B, Gemma-2-9B). Each source item yields four matched variants: human-written (HUMAN), direct LLM generation (Free-LLM), LLM-rewritten human text (H2L), and same-generator LLM-rewritten LLM text (LLM2L). We evaluated five detectors (FastDetectGPT, Binoculars-falcon-7b, RADAR, BERT-Defense, RoBERTa-Defense) at a strict 1%-false-positive operating point (TPR@1%FPR). FastDetectGPT and Binoculars-falcon-7b detected 91.2% and 93.5\% of direct LLM text, but only 30.8% and 15.1% of human text an LLM had rewritten, a drop of 60-78 percentage points. The same detectors retained 78.3% and 83.0% recall when LLM text was rewritten by the same model, a much smaller decline of 10-13 points. RADAR followed the same pattern (66.8% to 12.2%), while BERT-Defense and RoBERTa-Defense stayed below 3% recall across all regimes. These results show that detector performance measured on the conventional human-vs-LLM benchmark does not transfer to human-authored text revised by an LLM, even though the same detectors remain largely robust to LLM-only rewriting.

  • 2 authors
·
Jul 30 1

OmniRAS: Standardizing Foundation Model Training and Evaluation in Robot-Assisted Surgery

Few foundation models exist for robot-assisted surgery, partly because large robotic-surgery video corpora are difficult to assemble and existing models are evaluated mostly on laparoscopic benchmarks. Further, most existing models are evaluated on a small set of public benchmarks, mostly focused on laparoscopic surgery. We present OmniRAS, a family of 1B- and 2B-parameter V-JEPA-2.1 encoders for robot-assisted surgery, and detail their training. First, we release two densely annotated robotic-cholecystectomy datasets: OmniRAS-PR and a multi-label YT-Chole tool-verb-target task, the first triplet-style annotation for robotic cholecystectomy, together with splits, probe protocols, and an inter-rater study validating the shared phase ontology. Second, we document continued pretraining at up to 256 compute nodes with global batch 6,144 over 19 sources totaling approximately 2,650 hours of surgical video, 51% robotic, and analyze compute and data composition. Third, we evaluate against raw V-JEPA-2.1 and specialized surgical models on six tasks spanning triplet, phase, and step recognition, action segmentation, and detection, under frozen-encoder and final-four-block fine-tuning regimes. Across three seeds, this yields 254 downstream runs, including 109 with partial backbone fine-tuning. The best OmniRAS models achieve the strongest adapted results across all task families, while frozen differences are smaller.

  • 11 authors
·
Sep 8

Learning What to Forget: Distributional Unlearning for LLM Representation Spaces

Machine learning systems increasingly face the need to remove the influence of entire data domains, such as toxic language, harmful behavior, or topical content, rather than isolated records. Recent work formalizes this problem as distributional unlearning: selecting a subset of a forget domain whose removal moves the training distribution away from an unwanted population while preserving proximity to the desired one. However, existing analyses often impose parametric assumptions to obtain tractable selection rules. These assumptions may be poorly suited to high-dimensional language-model representations. We introduce Mamushi, a framework for non-parametric distributional unlearning that ranks forget examples using a probabilistic classifier whose Bayes-optimal logit equals the forget-to-retain log-density ratio (up to an additive class-prior constant). We show that thresholding the population log-density ratio yields the optimal fixed-budget selection rule for our removal--preservation objective and establish a non-asymptotic transfer guarantee relating score-estimation and threshold-calibration errors to degradation from the population-optimal selection rule. Our empirical evaluation spans real-world datasets on toxic-language removal and topical-domain removal regimes using different representations, with Mamushi achieving a more favorable removal--preservation trade-off than other baselines. Our work shows that Mamushi can serve as an efficient selection approach for downstream machine unlearning procedures, reducing the number of forget examples required to reach a fixed forgetting target.

  • 4 authors
·
Sep 29

HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMs

Hybrid-reasoning large language models (LLMs) expose explicit controls over reasoning effort, allowing users or systems to trade off answer quality against inference cost. However, existing methods for adaptive thinking-mode selection are typically evaluated under different models, datasets, and implementation assumptions, making it difficult to compare their practical behavior. We introduce HRBench, a unified evaluation framework for studying thinking-mode switching in hybrid-reasoning LLMs. HRBench organizes the design space along two axes: three switching strategy families, prompt-based selection, external routing, and speculative execution, and four training regimes, training-free, SFT, offline and online RL, yielding 12 controlled evaluation settings. We evaluate these settings across 6 LLMs, from Qwen3.5-2B to Kimi-K2.5-1.1T, and 5 reasoning benchmarks covering mathematics, science, and code, while reimplementing 12+ representative prior methods within the same pipeline. Our analysis characterizes how different switching strategies occupy distinct effectiveness-efficiency trade-off regions: prompt-based methods often provide favorable token-accuracy trade-offs, routing methods offer more stable cost reduction, and speculative methods tend to improve accuracy at higher token cost. We further find that training affects strategies differently, and that the preferred strategy varies with model scale and task domain. HRBench provides reference implementations and a unified evaluation platform to support more controlled research on efficient reasoning in hybrid-reasoning LLMs. Our data, code and repository are available at https://github.com/usail-hkust/HRBench.

tencent Tencent
·
May 26 4

What Makes Adversarial Examples Transfer Across Deepfake Detectors?

Deepfake detectors remain vulnerable to transfer-based black-box attacks, in which adversarial examples are generated on a source surrogate model and transferred to a target model, unknown to the attacker. Yet how source--target compatibility shapes attack success remains poorly understood. Prior studies evaluate limited detector pools and rarely disentangle architectural from training factors. We conduct a controlled evaluation of adversarial transferability across 60 detectors spanning six backbones, two pretraining regimes, and five training-data configurations, using two attack procedures: AutoAttack (AA) and the Carlini--Wagner attack with Expectation over Transformation (CW--EOT). Matched comparisons reveal significantly higher transfer when source and target share an exact backbone, architecture family, pretraining regime, or training data. This compatibility structure is attack-dependent: exact backbone compatibility has the largest effect under AA, whereas shared pretraining and training data have the largest effects under CW--EOT. When transfer is averaged across non-target sources, mean attack success rate (ASR) is 7.21% under AA and 19.52% under CW--EOT. By contrast, a multi-source oracle combining both attacks attains a \(64.48\%\) mean ASR after excluding exact backbone and training-data matches, showing that source averaging can substantially understate target vulnerability. We release 240,000 adversarially perturbed images, complete pairwise transfer results, detector configurations, and evaluation code. These findings establish source--target compatibility and source-model selection as central dimensions of credible transfer-based black-box robustness evaluation.

  • 3 authors
·
Sep 8

VP-Hype: A Hybrid Mamba-Transformer Framework with Visual-Textual Prompting for Hyperspectral Image Classification

Accurate classification of hyperspectral imagery (HSI) is often frustrated by the tension between high-dimensional spectral data and the extreme scarcity of labeled training samples. While hierarchical models like LoLA-SpecViT have demonstrated the power of local windowed attention and parameter-efficient fine-tuning, the quadratic complexity of standard Transformers remains a barrier to scaling. We introduce VP-Hype, a framework that rethinks HSI classification by unifying the linear-time efficiency of State-Space Models (SSMs) with the relational modeling of Transformers in a novel hybrid architecture. Building on a robust 3D-CNN spectral front-end, VP-Hype replaces conventional attention blocks with a Hybrid Mamba-Transformer backbone to capture long-range dependencies with significantly reduced computational overhead. Furthermore, we address the label-scarcity problem by integrating dual-modal Visual and Textual Prompts that provide context-aware guidance for the feature extraction process. Our experimental evaluation demonstrates that VP-Hype establishes a new state of the art in low-data regimes. Specifically, with a training sample distribution of only 2\%, the model achieves Overall Accuracy (OA) of 99.69\% on the Salinas dataset and 99.45\% on the Longkou dataset. These results suggest that the convergence of hybrid sequence modeling and multi-modal prompting provides a robust path forward for high-performance, sample-efficient remote sensing.

  • 7 authors
·
Mar 1

Adapting Without Gradients: Affine Statistics Transport and What Its Certificate Can Tell You

Test-time adaptation (TTA) typically assumes that model parameters can be updated at inference time. This assumption is restrictive for inference-only accelerators, frozen or third-party models, and memory-constrained deployments, and standard BatchNorm-based TTA configurations may also become inactive on architectures without BatchNorm. We study adaptation when the learned model must remain frozen. We introduce CASTER, a gradient-free method that stores source class statistics in a discriminative subspace, estimates a class-shared affine transformation from target-batch moments, and analytically transports the source class distributions before classification. CASTER requires no backward pass, optimizer state, or stored source feature bank. Across four backbones and seven datasets, it outperforms k-NN on identical frozen features in 27 of 28 backbone-dataset settings while retaining a median of 18x less state. Affine transport is not always reliable. On ImageNet-C, where batches contain only 64 samples for 1000 classes, unconditional transport loses 21.2 top-1 points. We therefore introduce an empirical residual-to-margin transportability certificate. Across 307 evaluation cells, every transport losing more than 10 points has certificate value above 3.9, although benign and destructive regimes are not perfectly separated. Gating converts an average -3.35-point effect of unconditional transport into a +1.69-point gain, and performance remains within 0.3 points of the best threshold over a broad threshold range. Finally, we show that this certificate is mechanism-specific: when applied to Tent, it accepts only 4.3% of updates and preserves 0.6% of Tent's available gain. These results position CASTER as a lightweight adaptation mechanism for frozen-model deployment, together with an explicit account of when its safety signal is informative and when it is not.

Talan Talan
·
Aug 30 2

SPEED-Bench: A Unified and Diverse Benchmark for Speculative Decoding

Speculative Decoding (SD) has emerged as a critical technique for accelerating Large Language Model (LLM) inference. Unlike deterministic system optimizations, SD performance is inherently data-dependent, meaning that diverse and representative workloads are essential for accurately measuring its effectiveness. Existing benchmarks suffer from limited task diversity, inadequate support for throughput-oriented evaluation, and a reliance on high-level implementations that fail to reflect production environments. To address this, we introduce SPEED-Bench, a comprehensive suite designed to standardize SD evaluation across diverse semantic domains and realistic serving regimes. SPEED-Bench offers a carefully curated Qualitative data split, selected by prioritizing semantic diversity across the data samples. Additionally, it includes a Throughput data split, allowing speedup evaluation across a range of concurrencies, from latency-sensitive low-batch settings to throughput-oriented high-load scenarios. By integrating with production engines like vLLM and TensorRT-LLM, SPEED-Bench allows practitioners to analyze system behaviors often masked by other benchmarks. We highlight this by quantifying how synthetic inputs overestimate real-world throughput, identifying batch-size dependent optimal draft lengths and biases in low-diversity data, and analyzing the caveats of vocabulary pruning in state-of-the-art drafters. We release SPEED-Bench to establish a unified evaluation standard for practical comparisons of SD algorithms.

nvidia NVIDIA
·
Feb 9 2

RealCADBench: Benchmarking Parametric CAD Modeling from Industrial Design Intents

Parametric computer-aided design (CAD) modeling is difficult to evaluate with a single metric. Existing CAD benchmarks often emphasize synthetic or CAD-native settings, limited input modalities, or executability and IoUs alone. We introduce RealCADBench, a benchmark for intent-to-program CAD modeling from real industrial design intents. It contains 12,632 tasks from 19 factory-automation categories and spans text descriptions, 2D engineering drawings, real product pictures, and rendered images for both Part and Assembly modeling. We report results on a 1,770-task evaluation slice: 1,745 Part tasks across four input regimes and RCB-Assm25, a 25-task assembly study used in every reported assembly comparison. Each method generates FreeCAD API Python, which a shared runtime executes to export the 3D model. We evaluate the exported model using executability, Solid IoU, Surface IoU, and a rubric-based visual-semantic identity Judge. Among the nine standalone frontier large models evaluated, no model leads all four metrics. Across six frontier-scale large models, executability ranges from 0.565 to 0.812, Solid IoU from 0.2841 to 0.5379, and Surface IoU from 0.112 to 0.217 across the four Part regimes. The highest regime-balanced composite comes from a different model than the leaders on the four component metrics. On RCB-Assm25, Codex with GPT-5.5 improves executability and both IoU metrics over standalone GPT-5.5, but lowers the Judge score by 6.98 percentage points, leaving GPT-5.5 as the Judge leader. We also observe recurring failure modes, most notably missing fine structures, loss of part identity, and incorrect assembly placement. These results show that execution alone is insufficient to characterize realistic CAD modeling and that frontier models and agents differ substantially across executability, IoUs, and visual-semantic identity.

  • 11 authors
·
Sep 3

Accurate Estimation of Mutual Information in High Dimensional Data

Mutual information (MI) is a fundamental measure of statistical dependence between two variables, yet accurate estimation from finite data remains notoriously difficult. No estimator is universally reliable, and common approaches fail in the high-dimensional, undersampled regimes typical of modern experiments. Recent machine learning-based estimators show promise, but their accuracy depends sensitively on dataset size, structure, and hyperparameters, with no accepted tests to detect failures. We close these gaps through a systematic evaluation of classical and neural MI estimators across standard benchmarks and new synthetic datasets tailored to challenging high-dimensional, undersampled regimes. We contribute: (i) a practical protocol for reliable MI estimation with explicit checks for statistical consistency; (ii) confidence intervals (error bars around estimates) that existing neural MI estimator do not provide; and (iii) a new class of probabilistic critics designed for high-dimensional, high-information settings. We demonstrate the effectiveness of our protocol with computational experiments, showing that it consistently matches or surpasses existing methods while uniquely quantifying its own reliability. We show that reliable MI estimation is sometimes achievable even in severely undersampled, high-dimensional datasets, provided they admit accurate low-dimensional representations. This broadens the scope of applicability of neural MI estimators and clarifies when such estimators can be trusted.

  • 3 authors
·
May 30, 2025

Dynamic Delayed Tree Expansion For Improved Multi-Path Speculative Decoding

Multi-path speculative decoding accelerates lossless sampling from a target model by using a cheaper draft model to generate a draft tree of tokens, and then applies a verification algorithm that accepts a subset of these. While prior work has proposed various verification algorithms for i.i.d rollouts, their relative performance under matched settings remains unclear. In this work, we firstly present a systematic evaluation of verification strategies across model families, tasks, and sampling regimes, and find that Traversal Verification dominates consistently, with OT-based methods lagging far behind. Our analysis uncovers that this occurs because OT-based methods achieve high multi-token acceptance near the root of the draft tree, while multi-token gains are most impactful deeper in the draft tree, where draft and target distributions diverge. Based on this insight, we propose delayed tree expansion, which drafts a partial single path, delaying the i.i.d. branching point. We show that delayed tree expansion preserves the target distribution and improves on root-node i.i.d rollouts. Further, we develop a dynamic neural selector that estimates the expected block efficiency of optimal-transport-based verification methods from draft and target features, enabling context-dependent expansion decisions. Our neural selector allows OT-based methods like SpecInfer to outperform Traversal Verification for the first time, achieving 5% higher average throughput across a wide range of models, datasets, and sampling settings.

  • 4 authors
·
Feb 18

KARL: Knowledge Agents via Reinforcement Learning

We present a system for training enterprise search agents via reinforcement learning that achieves state-of-the-art performance across a diverse suite of hard-to-verify agentic search tasks. Our work makes four core contributions. First, we introduce KARLBench, a multi-capability evaluation suite spanning six distinct search regimes, including constraint-driven entity search, cross-document report synthesis, tabular numerical reasoning, exhaustive entity retrieval, procedural reasoning over technical documentation, and fact aggregation over internal enterprise notes. Second, we show that models trained across heterogeneous search behaviors generalize substantially better than those optimized for any single benchmark. Third, we develop an agentic synthesis pipeline that employs long-horizon reasoning and tool use to generate diverse, grounded, and high-quality training data, with iterative bootstrapping from increasingly capable models. Fourth, we propose a new post-training paradigm based on iterative large-batch off-policy RL that is sample efficient, robust to train-inference engine discrepancies, and naturally extends to multi-task training with out-of-distribution generalization. Compared to Claude 4.6 and GPT 5.2, KARL is Pareto-optimal on KARLBench across cost-quality and latency-quality trade-offs, including tasks that were out-of-distribution during training. With sufficient test-time compute, it surpasses the strongest closed models. These results show that tailored synthetic data in combination with multi-task reinforcement learning enables cost-efficient and high-performing knowledge agents for grounded reasoning.

databricks Databricks
·
Mar 5 1

AnyGoal: On World Model Generalization Beyond Expert Rollouts

The goal of a world model is to capture the dynamics of an environment. This has the great appeal that it allows agents to plan and simulate a course of actions with the help of a world model. For these uses to be reliable, however, the learned dynamics must behave well beyond the specific state-action combinations observed in expert rollouts. Despite this, the relationship between data coverage, dynamics generalization, and downstream planning performance remains underexplored. We study this relationship through the lens of compositional generalization: can a world model recombine familiar state and action factors into valid dynamics? To answer this, we introduce AnyGoal, a benchmark of goal-shifted task variants built from widely used world-model environments, together with an open-source library. Each task is designed to be within reach of a world model that has learned the underlying dynamics: individual state factors and actions remain within their training support, but their combinations are underrepresented in expert trajectories. Across multiple environments and world model families, we find that current models consistently exhibit failures under these recombinations. By examining dynamics and representation failures across model families, interaction regimes, and evaluation metrics, we study where current world models generalize and where challenges remain for goal-shifted prediction and planning. Together, this framework, benchmark, and analysis provide a basis for evaluating progress toward world models that generalize to new planning tasks.

FlexGuard: Continuous Risk Scoring for Strictness-Adaptive LLM Content Moderation

Ensuring the safety of LLM-generated content is essential for real-world deployment. Most existing guardrail models formulate moderation as a fixed binary classification task, implicitly assuming a fixed definition of harmfulness. In practice, enforcement strictness - how conservatively harmfulness is defined and enforced - varies across platforms and evolves over time, making binary moderators brittle under shifting requirements. We first introduce FlexBench, a strictness-adaptive LLM moderation benchmark that enables controlled evaluation under multiple strictness regimes. Experiments on FlexBench reveal substantial cross-strictness inconsistency in existing moderators: models that perform well under one regime can degrade substantially under others, limiting their practical usability. To address this, we propose FlexGuard, an LLM-based moderator that outputs a calibrated continuous risk score reflecting risk severity and supports strictness-specific decisions via thresholding. We train FlexGuard via risk-alignment optimization to improve score-severity consistency and provide practical threshold selection strategies to adapt to target strictness at deployment. Experiments on FlexBench and public benchmarks demonstrate that FlexGuard achieves higher moderation accuracy and substantially improved robustness under varying strictness. We release the source code and data to support reproducibility.

  • 4 authors
·
Feb 26

FOS: A Large-Scale Temporal Graph Benchmark for Scientific Interdisciplinary Link Prediction

Interdisciplinary scientific breakthroughs mostly emerge unexpectedly, and forecasting the formation of novel research fields remains a major challenge. We introduce FOS (Future Of Science), a comprehensive time-aware graph-based benchmark that reconstructs annual co-occurrence graphs of 65,027 research sub-fields (spanning 19 general domains) over the period 1827-2024. In these graphs, edges denote the co-occurrence of two fields in a single publication and are timestamped with the corresponding publication year. Nodes are enriched with semantic embeddings, and edges are characterized by temporal and topological descriptors. We formulate the prediction of new field-pair linkages as a temporal link-prediction task, emphasizing the "first-time" connections that signify pioneering interdisciplinary directions. Through extensive experiments, we evaluate a suite of state-of-the-art temporal graph architectures under multiple negative-sampling regimes and show that (i) embedding long-form textual descriptions of fields significantly boosts prediction accuracy, and (ii) distinct model classes excel under different evaluation settings. Case analyses show that top-ranked link predictions on FOS align with field pairings that emerge in subsequent years of academic publications. We publicly release FOS, along with its temporal data splits and evaluation code, to establish a reproducible benchmark for advancing research in predicting scientific frontiers.

  • 7 authors
·
Nov 23, 2025

When Network Architecture Meets Physics: Deep Operator Learning for Coupled Multiphysics

Scientific applications increasingly demand real-time surrogate models that can capture the behavior of strongly coupled multiphysics systems driven by multiple input functions, such as in thermo-mechanical and electro-thermal processes. While neural operator frameworks, such as Deep Operator Networks (DeepONets), have shown considerable success in single-physics settings, their extension to multiphysics problems remains poorly understood. In particular, the challenge of learning nonlinear interactions between tightly coupled physical fields has received little systematic attention. This study addresses a foundational question: should the architectural design of a neural operator reflect the strength of physical coupling it aims to model? To answer this, we present the first comprehensive, architecture-aware evaluation of DeepONet variants across three regimes: single-physics, weakly coupled, and strongly coupled multiphysics systems. We consider a reaction-diffusion equation with dual spatial inputs, a nonlinear thermo-electrical problem with bidirectional coupling through temperature-dependent conductivity, and a viscoplastic thermo-mechanical model of steel solidification governed by transient phase-driven interactions. Two operator-learning frameworks, the classical DeepONet and its sequential GRU-based extension, S-DeepONet, are benchmarked using both single-branch and multi-branch (MIONet-style) architectures. Our results demonstrate that architectural alignment with physical coupling is crucial: single-branch networks significantly outperform multi-branch counterparts in strongly coupled settings, whereas multi-branch encodings offer advantages for decoupled or single-physics problems. Once trained, these surrogates achieve full-field predictions up to 1.8e4 times faster than high-fidelity finite-element solvers, without compromising solution accuracy.

  • 6 authors
·
Jul 3, 2025

MacroLens: A Multi-Task Benchmark for Contextual Financial Reasoning under Macroeconomic Scenarios

Financial decision-making is contextual: forecasting prices, valuing companies, and assessing event exposure weigh price history, accounting fundamentals, macroeconomic regime, and contemporaneous text. A benchmark over these four signals is hard to build because finance violates four assumptions of time-series evaluation: text must be gated by its publication date to prevent look-ahead, quarterly fundamentals are reported with a one- to ninety-day lag, filing text is partly redundant with the numerical statement fields it accompanies, and macroeconomic regimes leak across calendar splits. No public benchmark addresses all four signals jointly. MacroLens covers 4,416 U.S. small- and micro-cap equities over 2021-2026. Seven tasks share one point-in-time panel of prices, 46.8M XBRL accounting facts, 53 macroeconomic series, 295,860 SEC filings, and 215,882 news articles, plus a scenario layer of 1,130 macroeconomic events across 49 types automatically detected and rendered as natural language. Tasks span contextual forecasting, public and private valuation, statement generation from fundamentals and descriptions, scenario-conditioned returns, and real-estate valuation. We evaluate 19 methods across six families spanning naive heuristics through time-series foundation models, fine-tuned LLM-based time-series models, and zero-shot large language models (LLMs), plus a five-step feature-context ablation on two frontier LLMs and a gradient-boosted baseline. MacroLens is released at https://huggingface.co/datasets/DeepAuto-AI/MacroLens.

  • 5 authors
·
Jun 22

Advancing Open and Reproducible Relational Learning: RelArena-$α$, TabPFN-Rel and RPI

This first release of Prior Labs in relational learning shows our continued commitment to open science. We open-source three pieces of software that we expect to accelerate research in the field towards meaningful real-world impact. We aim to steer further development based on feedback from, and in collaboration with, the community. Given the early stage of development, our α-release targets researchers and early-adopting practitioners. Over the past years, a variety of datasets and tasks for relational learning have emerged, but the community has not converged on a reliable, reproducible way to compare different methods on these tasks. Our α-release, RelArena-α, provides a unified framework for running and comparing baselines on RelBench v1 by standardizing data loading, evaluation protocols, tuning regimes, and support for systems with custom tuning, inspired by established tabular benchmarks such as TabArena. We plan to work with the research community to further develop RelArena-α into a catalyst for progress in the relational learning community. We release the initial version of TabPFN-Rel, a purpose-built relational harness for TabPFN-3. Currently ranked first among models on RelArena-α, TabPFN-Rel makes key improvements upon RDBLearn. Beyond its ranking, TabPFN-Rel serves as a strong baseline, adding to the growing evidence that flattening a relational database into a single table remains competitive with specialized relational architectures on real-world tasks. To facilitate adoption of relational learning methods in research and industry, we release an initial α-version of our Relational Predictive Interface, RPI, an open-source, model-agnostic interface that enables early adopters to easily define problems on new databases and apply any model implemented in RelArena-α, including TabPFN-Rel, to these problems.

Prior-Labs Prior Labs
·
Aug 16 2

CroCoMAE: Cross-modal self-supervised collaborative representation learning for SAR-optical remote sensing data

Remote sensing increasingly captures the Earth's surface through complementary sensors such as SAR and optical, yet ground-truth annotations remain scarce, limiting supervised approaches. Self-supervised learning (SSL) addresses the label scarcity, and cross-modal SSL further exploits the complementarity between co-located sensors, but existing cross-modal frameworks align modalities mostly through global image-level representations or treat them as largely redundant, leaving their fine-grained complementary structure underexploited. We propose CroCoMAE, a cross-modal SSL framework for SAR-optical remote sensing that unifies contrastive learning and masked image modeling within a ViT-based architecture, operating at two complementary levels. At the image level, an intra-modality contrastive loss on the class token yields globally discriminative, augmentation-invariant representations. At the patch level, masked image modeling drives fine-grained per-modality representations, while a cross-modal contrastive objective aligns a shared subspace across SAR and optical patch tokens, leaving the complementary subspace unconstrained to retain modality-specific information. On two SAR-optical benchmarks covering scene classification and semantic segmentation, CroCoMAE improves over uni-modal SSL baselines and outperforms state-of-the-art cross-modal frameworks in most configurations, with gains in both single-modality and multi-modal evaluation across varying label regimes and the largest improvements on semantic segmentation.

  • 6 authors
·
Nov 9

QuitoBench: A High-Quality Open Time Series Forecasting Benchmark

Time series forecasting is critical across finance, healthcare, and cloud computing, yet progress is constrained by a fundamental bottleneck: the scarcity of large-scale, high-quality benchmarks. To address this gap, we introduce QuitoBench, a regime-balanced benchmark for time series forecasting with coverage across eight trendtimesseasonalitytimesforecastability (TSF) regimes, designed to capture forecasting-relevant properties rather than application-defined domain labels. The benchmark is built upon Quito, a billion-scale time series corpus of application traffic from Alipay spanning nine business domains. Benchmarking 10 models from deep learning, foundation models, and statistical baselines across 232,200 evaluation instances, we report four key findings: (i) a context-length crossover where deep learning models lead at short context (L=96) but foundation models dominate at long context (L ge 576); (ii) forecastability is the dominant difficulty driver, producing a 3.64 times MAE gap across regimes; (iii) deep learning models match or surpass foundation models at 59 times fewer parameters; and (iv) scaling the amount of training data provides substantially greater benefit than scaling model size for both model families. These findings are validated by strong cross-benchmark and cross-metric consistency. Our open-source release enables reproducible, regime-aware evaluation for time series forecasting research.

  • 10 authors
·
Mar 26 3

Combining Synthetic and Real Data for Low-Resource Historical OCR: A Manchu Case Study

Manchu, now critically endangered, was one of the principal languages of the Qing empire (1636-1912), and its extensive archival record is increasingly digitized but remains difficult to search and analyze at scale. Previous work showed that vision-language models (VLMs) trained only on synthetic Manchu word images can reach 87.4% word accuracy on real Qing manuscripts and prints, leaving a substantial synthetic-to-real gap. This study examines how synthetic and real historical training data should be combined for low-resource OCR. Using 60,000 synthetic and 20,306 real historical word images, we evaluate three pretrained VLMs and a compact convolutional recurrent neural network (CRNN) under four regimes: synthetic-only, real-only, joint synthetic-real, and sequential synthetic-to-real training, following a common checkpoint-selection and archival evaluation protocol. Introducing real training images raises the leading configurations to between 95.09% and 96.28% word accuracy, while no synthetic-only configuration exceeds 87.92%. Synthetic supplementation substantially improves all three VLMs, whereas its marginal effect for the CRNN is sensitive to the training objective. Joint and sequential training yield broadly similar archival accuracy under the tested practical pipelines. A compact CRNN also reaches the leading performance range once real images are available, showing that model scale alone does not determine recognition accuracy. Finally, complementary errors among strong recognizers allow voting to raise accuracy to 98.27% without additional training, while an eighteenth-century Manchu dictionary provides a principled rule for adjudicating disagreements.

  • 2 authors
·
Sep 9

Cookie-Bench: Continuous On-screen Key Interaction Evaluation for Web Generation

Front-end web code has become a core product surface for every frontier LLM release, yet evaluating these interactive applications at development speed remains costly because human-judged leaderboards like Arena do not scale. Existing automated proxies typically lean on reference implementations, test suites, or rigid checklists, and tend to miss the reasoned synthesis a human reviewer performs over a live session. We articulate a new evaluation regime that is simultaneously reference-free, autonomously driven, and holistically reasoned, and instantiate it through two artifacts. \dataname is an 11-domain, 54-leaf, 1,000-query WebDev benchmark spanning both static-presentation and interactive-application tasks, balanced across three difficulty tiers and three target-language groups, with briefs rewritten to resist recall from circulated prompts. \framename, grounded in Flavell's metacognitive monitoring, separates evidence accumulation from judgment across three stages: Static Perception forms a first impression from passive observation; Agent-Driven Interaction explores the application autonomously while capturing continuous screen video, audio, and per-step screenshots; Dynamic Scoring issues holistic functionality and aesthetics verdicts with structured failure attribution only after the evidence chain is complete. On \dataname, \framename aligns closely with expert human ratings while surfacing substantial headroom across 13 frontier LLMs on interactive web generation. \noindenthttps://anonymous.4open.science/r/Cookie-3CE/

  • 11 authors
·
May 30

Cultural Evolution of Cooperation among LLM Agents

Large language models (LLMs) provide a compelling foundation for building generally-capable AI agents. These agents may soon be deployed at scale in the real world, representing the interests of individual humans (e.g., AI assistants) or groups of humans (e.g., AI-accelerated corporations). At present, relatively little is known about the dynamics of multiple LLM agents interacting over many generations of iterative deployment. In this paper, we examine whether a "society" of LLM agents can learn mutually beneficial social norms in the face of incentives to defect, a distinctive feature of human sociality that is arguably crucial to the success of civilization. In particular, we study the evolution of indirect reciprocity across generations of LLM agents playing a classic iterated Donor Game in which agents can observe the recent behavior of their peers. We find that the evolution of cooperation differs markedly across base models, with societies of Claude 3.5 Sonnet agents achieving significantly higher average scores than Gemini 1.5 Flash, which, in turn, outperforms GPT-4o. Further, Claude 3.5 Sonnet can make use of an additional mechanism for costly punishment to achieve yet higher scores, while Gemini 1.5 Flash and GPT-4o fail to do so. For each model class, we also observe variation in emergent behavior across random seeds, suggesting an understudied sensitive dependence on initial conditions. We suggest that our evaluation regime could inspire an inexpensive and informative new class of LLM benchmarks, focussed on the implications of LLM agent deployment for the cooperative infrastructure of society.

  • 2 authors
·
Dec 13, 2024

Bridging the Gap Between Computational Photography and Visual Recognition

What is the current state-of-the-art for image restoration and enhancement applied to degraded images acquired under less than ideal circumstances? Can the application of such algorithms as a pre-processing step to improve image interpretability for manual analysis or automatic visual recognition to classify scene content? While there have been important advances in the area of computational photography to restore or enhance the visual quality of an image, the capabilities of such techniques have not always translated in a useful way to visual recognition tasks. Consequently, there is a pressing need for the development of algorithms that are designed for the joint problem of improving visual appearance and recognition, which will be an enabling factor for the deployment of visual recognition tools in many real-world scenarios. To address this, we introduce the UG^2 dataset as a large-scale benchmark composed of video imagery captured under challenging conditions, and two enhancement tasks designed to test algorithmic impact on visual quality and automatic object recognition. Furthermore, we propose a set of metrics to evaluate the joint improvement of such tasks as well as individual algorithmic advances, including a novel psychophysics-based evaluation regime for human assessment and a realistic set of quantitative measures for object recognition performance. We introduce six new algorithms for image restoration or enhancement, which were created as part of the IARPA sponsored UG^2 Challenge workshop held at CVPR 2018. Under the proposed evaluation regime, we present an in-depth analysis of these algorithms and a host of deep learning-based and classic baseline approaches. From the observed results, it is evident that we are in the early days of building a bridge between computational photography and visual recognition, leaving many opportunities for innovation in this area.

  • 24 authors
·
Jan 27, 2019

Summarization Bias: The Directional Collapse of Objective Projection into Told-Mode Labels in Large Language Models --- A Conceptual Framework and Registered Test Protocol

This paper introduces and operationalizes summarization bias: a proposed systematic tendency of large language models (LLMs) to represent narrative meaning as an abstract summary label rather than as the reconstructable inferential structure that produces it. Within the Bulut Doctrine, narrative effect is theorized along a told-shown axis: in told mode, emotional and informational content is declared explicitly and requires little reader reconstruction; in shown mode, that content is suppressed at the surface and must be reconstructed from physical cues and indirection (Objective Projection). Shown mode is the higher-load condition the doctrine is designed to measure. The claim is that LLMs fail along this axis in a specific direction. Summarization bias is hypothesized to operate in two regimes: (i) a generative regime, in which a model asked to render an emotion through Objective Projection defaults to declaring it instead; and (ii) an evaluative regime, in which a model judging narrative quality rewards told-mode explicitness and under-detects shown-mode suppression. The evaluative regime is the more consequential, since LLMs increasingly serve as judges and reward models, and a directional bias toward told mode would impose a selection pressure degrading prose toward flat declaration. This report does not claim the bias is validated. It defines the construct, situates it against LLM-as-judge biases, rereads a completed independent reliability study as directional evidence consistent with it, and pre-registers a two-regime test with decision rules under which the construct would be abandoned.

  • 1 authors
·
Sep 17

EvalAgent: Discovering Implicit Evaluation Criteria from the Web

Evaluation of language model outputs on structured writing tasks is typically conducted with a number of desirable criteria presented to human evaluators or large language models (LLMs). For instance, on a prompt like "Help me draft an academic talk on coffee intake vs research productivity", a model response may be evaluated for criteria like accuracy and coherence. However, high-quality responses should do more than just satisfy basic task requirements. An effective response to this query should include quintessential features of an academic talk, such as a compelling opening, clear research questions, and a takeaway. To help identify these implicit criteria, we introduce EvalAgent, a novel framework designed to automatically uncover nuanced and task-specific criteria. EvalAgent first mines expert-authored online guidance. It then uses this evidence to propose diverse, long-tail evaluation criteria that are grounded in reliable external sources. Our experiments demonstrate that the grounded criteria produced by EvalAgent are often implicit (not directly stated in the user's prompt), yet specific (high degree of lexical precision). Further, EvalAgent criteria are often not satisfied by initial responses but they are actionable, such that responses can be refined to satisfy them. Finally, we show that combining LLM-generated and EvalAgent criteria uncovers more human-valued criteria than using LLMs alone.

  • 6 authors
·
Apr 21, 2025

Quantifying Variance in Evaluation Benchmarks

Evaluation benchmarks are the cornerstone of measuring capabilities of large language models (LLMs), as well as driving progress in said capabilities. Originally designed to make claims about capabilities (or lack thereof) in fully pretrained models, evaluation benchmarks are now also extensively used to decide between various training choices. Despite this widespread usage, we rarely quantify the variance in our evaluation benchmarks, which dictates whether differences in performance are meaningful. Here, we define and measure a range of metrics geared towards measuring variance in evaluation benchmarks, including seed variance across initialisations, and monotonicity during training. By studying a large number of models -- both openly available and pretrained from scratch -- we provide empirical estimates for a variety of variance metrics, with considerations and recommendations for practitioners. We also evaluate the utility and tradeoffs of continuous versus discrete performance measures and explore options for better understanding and reducing this variance. We find that simple changes, such as framing choice tasks (like MMLU) as completion tasks, can often reduce variance for smaller scale (sim7B) models, while more involved methods inspired from human testing literature (such as item analysis and item response theory) struggle to meaningfully reduce variance. Overall, our work provides insights into variance in evaluation benchmarks, suggests LM-specific techniques to reduce variance, and more generally encourages practitioners to carefully factor in variance when comparing models.

  • 8 authors
·
Jun 14, 2024

Magic Words or Methodical Work? Challenging Conventional Wisdom in LLM-Based Political Text Annotation

Political scientists are rapidly adopting large language models (LLMs) for text annotation, yet the sensitivity of annotation results to implementation choices remains poorly understood. Most evaluations test a single model or configuration; how model choice, model size, learning approach, and prompt style interact, and whether popular "best practices" survive controlled comparison, are largely unexplored. We present a controlled evaluation of these pipeline choices, testing six open-weight models across four political science annotation tasks under identical quantisation, hardware, and prompt-template conditions. Our central finding is methodological: interaction effects dominate main effects, so seemingly reasonable pipeline choices can become consequential researcher degrees of freedom. No single model, prompt style, or learning approach is uniformly superior, and the best-performing model varies across tasks. Two corollaries follow. First, model size is an unreliable guide both to cost and to performance: cross-family efficiency differences are so large that some larger models are less resource-intensive than much smaller alternatives, while within model families mid-range variants often match or exceed larger counterparts. Second, widely recommended prompt engineering techniques yield inconsistent and sometimes negative effects on annotation performance. We use these benchmark results to develop a validation-first framework - with a principled ordering of pipeline decisions, guidance on prompt freezing and held-out evaluation, reporting standards, and open-source tools - to help researchers navigate this decision space transparently.

  • 5 authors
·
Mar 27

What AI Red-Team Evaluations Can and Cannot Prove

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.

apisec apisec
·
Jul 22 2

Deep Research, Shallow Evaluation: A Case Study in Meta-Evaluation for Long-Form QA Benchmarks

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.

  • 12 authors
·
Mar 5

HANDBOOK.md: A Benchmark for Long-Context Agentic Instruction Following

Language-model agents are increasingly deployed under standing instructions: a system prompt, a policy file, or a skills document is placed in context, and the agent is trusted to let that document govern every action that follows. Existing benchmarks rarely test this deployment pattern directly; they measure whether an agent can complete a task, not whether a long, binding policy document constrains its behavior over an extended tool-use horizon. We present HANDBOOK_md, a benchmark of 65 agentic tasks modeled on how employees follow company handbooks. Each task places an agent in a self-contained company environment (a file workspace with mock email, chat, calendar, issue-tracking, and commerce services exposed over the Model Context Protocol) and instructs it to carry out routine professional work governed by an expert-written standard operating procedure of 20-124 pages. Tasks span five domains (finance, medical billing, insurance, logistics, and HR) and 10 fictional companies. To resist memorization, every task modifies one of 10 base handbooks, altering the specific rules and thresholds on which grading depends, so no two tasks share the same set of policies. Grading is fully deterministic: each task carries a rubric of programmatic criteria (824 in total) that check both that required actions occurred and that prohibited actions did not. Under strict grading, where a trial passes only if every criterion is satisfied, the strongest evaluated model passes 36.2% of trials, and most frontier models remain below 25%. Failures follow consistent patterns: agents let a plausible but unauthorized in-environment request override the standing policy, perform a required check and then act against its result, lose rule details over long horizons, and report compliance they did not achieve. We release the tasks, environments, and evaluation harness.

  • 7 authors
·
Aug 2

Towards Multi-Method Evaluation of AI in Health: Integrating Benchmarks, Autoraters, and Expert Evaluation

General-purpose generative AI (GenAI) models are being rapidly adopted in healthcare, but they are challenging to assess with traditional evaluation frameworks. These conceptual, methodological, and practical challenges parallel decades of work in evaluating clinician and health system quality and safety. We describe the adaptation of methods from several fields, combined with emerging technical approaches, into a structured, multi-method evaluation framework designed to give well-rounded assessments of GenAI in health at scale. The approach first anchors evaluation to Critical User Journeys (CUJs) to help prioritize out of a large choice space, and maps these to a phased AI system lifecycle (base model development, product development, and post-deployment monitoring), since evaluation at each stage requires different methodological approaches. In this context, the approach then integrates three distinct evaluation methods (closed-question benchmarks, LLM-as-a-judge autoraters, and expert human review) across three core domains: accuracy, safety, and usefulness. Taken together, this approach provides a strategic framework for bridging the gap between benchmarking results and performance in the real world with the goal of supporting decision-making and accountability across development, procurement, and deployment for product teams, health systems, and regulators.

  • 33 authors
·
Nov 3

Zero-shot reasoning for simulating scholarly peer-review

The scholarly publishing ecosystem faces a dual crisis of unmanageable submission volumes and unregulated AI, creating an urgent need for new governance models to safeguard scientific integrity. The traditional human-only peer review regime lacks a scalable, objective benchmark, making editorial processes opaque and difficult to audit. Here we investigate a deterministic simulation framework that provides the first stable, evidence-based standard for evaluating AI-generated peer review reports. Analyzing 352 peer-review simulation reports, we identify consistent system state indicators that demonstrate its reliability. First, the system is able to simulate calibrated editorial judgment, with 'Revise' decisions consistently forming the majority outcome (>50%) across all disciplines, while 'Reject' rates dynamically adapt to field-specific norms, rising to 45% in Health Sciences. Second, it maintains unwavering procedural integrity, enforcing a stable 29% evidence-anchoring compliance rate that remains invariant across diverse review tasks and scientific domains. These findings demonstrate a system that is predictably rule-bound, mitigating the stochasticity of generative AI. For the scientific community, this provides a transparent tool to ensure fairness; for publishing strategists, it offers a scalable instrument for auditing workflows, managing integrity risks, and implementing evidence-based governance. The framework repositions AI as an essential component of institutional accountability, providing the critical infrastructure to maintain trust in scholarly communication.

  • 1 authors
·
Oct 2, 2025

The Evaluation Protocol Determines the Result: An Independent Reproduction of LeWorldModel on TwoRoom

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.

  • 1 authors
·
Aug 9

Large Language Models are not Fair Evaluators

In this paper, we uncover a systematic bias in the evaluation paradigm of adopting large language models~(LLMs), e.g., GPT-4, as a referee to score and compare the quality of responses generated by candidate models. We find that the quality ranking of candidate responses can be easily hacked by simply altering their order of appearance in the context. This manipulation allows us to skew the evaluation result, making one model appear considerably superior to the other, e.g., Vicuna-13B could beat ChatGPT on 66 over 80 tested queries with ChatGPT as an evaluator. To address this issue, we propose a calibration framework with three simple yet effective strategies: 1) Multiple Evidence Calibration, which requires the evaluator model to generate multiple evaluation evidence before assigning ratings; 2) Balanced Position Calibration, which aggregates results across various orders to determine the final score; 3) Human-in-the-Loop Calibration, which introduces a balanced position diversity entropy to measure the difficulty of each example and seeks human assistance when needed. We also manually annotate the "win/tie/lose" outcomes of responses from ChatGPT and Vicuna-13B in the Vicuna Benchmark's question prompt, and extensive experiments demonstrate that our approach successfully mitigates evaluation bias, resulting in closer alignment with human judgments. We release our code and human annotation at https://github.com/i-Eval/FairEval to facilitate future research.

  • 10 authors
·
May 29, 2023

CRAFT: Clustering Rubrics to Diagnose Weak LLM Capabilities and Generate Targeted Fine-Tuning Data

Evaluations should do more than measure a models current performance. They should tell us what to fix for the next model iteration and provide a way to generate targeted post training data. Most evaluation pipelines identify weak examples, topics, or categories, but they leave the underlying capability failure implicit: they say where a model fails, not why. We introduce CRAFT, a method that converts any rubric based evaluation dataset into a model specific diagnosis of weak capabilities. CRAFT treats each grading criterion as a capability probe: it extracts a capability description from every prompt rubric pair, clusters these descriptions into a hierarchical capability tree, scores the target model at every node, and selects low performing nodes dynamically across tree levels, at the granularity where each failure is clearest. The selected weak capabilities then direct the generation of targeted supervised finetuning data. Holding the data generation, finetuning, and evaluation setup fixed, we compare CRAFT against prompt level EvalTree clustering and untargeted random generation on four open source models, two professional domains (finance and legal), and 13 held out benchmarks disjoint from the diagnostic data. CRAFT achieves the strongest finance domain average for all four models under repeated temperature decoding; on legal domain, it is strongest for three of four models and remains within the decoding variance bands of the best baseline on the fourth. Diagnosing weaknesses at the level of rubric criteria, rather than prompts or categories, thus yields both a sharper picture of what a model cannot do and measurably better models after finetuning on that diagnosis.

  • 6 authors
·
Jul 16

Binarization Flattens the Score Space

Large language model (LLM) judges are often used as rewards to train policies on objectives that deterministic verifiers cannot capture. However, these rewards are often collapsed to pass/fail ({0, 1}), which reports the verdict but not how well a response met each criterion. We model each pass/fail verdict as a score on an unreported scale, compared with one cutoff. A stretch of that scale moves every score proportionally toward or away from the cutoff, but never across it, so no verdict changes. A policy is therefore free to apply any stretch without changing anything the panel reports. Under a joint-Gaussian model, a third grade adds a second threshold and removes this affine stretch ambiguity. On MATH and SciBench outputs from one seven-criterion judge, all 14 constructed criterionwise stretches were invisible after binarization but visible with three grades. At n=1{,}024, a test given both population laws had at least 96.5% power at a 1.5times stress. Retaining grades closes one blind spot created by binarization, but verdicts alone remain insufficient as some changes are still indistinguishable from genuine improvement. These include arbitrary within-grade changes and fixed-covariance, loading-aligned mean shifts -- the signature of a sycophancy-shaped lift the panel reads as competence. The shared-factor reference approximation fit MATH and SciBench but not HealthBench, delineating its empirical scope. We recommend keeping at least three grades (for example, asking the judge whether each criterion is fully, partially, or not met and rewarding {0, 0.5, 1}), and externally validating gains along the remaining direction, which no finer scale removes.

  • 1 authors
·
Sep 16

Evaluating the Social Impact of Generative AI Systems in Systems and Society

Generative AI systems across modalities, ranging from text (including code), image, audio, and video, have broad social impacts, but there is no official standard for means of evaluating those impacts or for which impacts should be evaluated. In this paper, we present a guide that moves toward a standard approach in evaluating a base generative AI system for any modality in two overarching categories: what can be evaluated in a base system independent of context and what can be evaluated in a societal context. Importantly, this refers to base systems that have no predetermined application or deployment context, including a model itself, as well as system components, such as training data. Our framework for a base system defines seven categories of social impact: bias, stereotypes, and representational harms; cultural values and sensitive content; disparate performance; privacy and data protection; financial costs; environmental costs; and data and content moderation labor costs. Suggested methods for evaluation apply to listed generative modalities and analyses of the limitations of existing evaluations serve as a starting point for necessary investment in future evaluations. We offer five overarching categories for what can be evaluated in a broader societal context, each with its own subcategories: trustworthiness and autonomy; inequality, marginalization, and violence; concentration of authority; labor and creativity; and ecosystem and environment. Each subcategory includes recommendations for mitigating harm.

  • 18 authors
·
Jun 9, 2023

SHARP: Social Harm Analysis via Risk Profiles for Measuring Inequities in Large Language Models

Large language models (LLMs) are increasingly deployed in high-stakes domains, where rare but severe failures can result in irreversible harm. However, prevailing evaluation benchmarks often reduce complex social risk to mean-centered scalar scores, thereby obscuring distributional structure, cross-dimensional interactions, and worst-case behavior. This paper introduces Social Harm Analysis via Risk Profiles (SHARP), a framework for multidimensional, distribution-aware evaluation of social harm. SHARP models harm as a multivariate random variable and integrates explicit decomposition into bias, fairness, ethics, and epistemic reliability with a union-of-failures aggregation reparameterized as additive cumulative log-risk. The framework further employs risk-sensitive distributional statistics, with Conditional Value at Risk (CVaR95) as a primary metric, to characterize worst-case model behavior. Application of SHARP to eleven frontier LLMs, evaluated on a fixed corpus of n=901 socially sensitive prompts, reveals that models with similar average risk can exhibit more than twofold differences in tail exposure and volatility. Across models, dimension-wise marginal tail behavior varies systematically across harm dimensions, with bias exhibiting the strongest tail severities, epistemic and fairness risks occupying intermediate regimes, and ethical misalignment consistently lower; together, these patterns reveal heterogeneous, model-dependent failure structures that scalar benchmarks conflate. These findings indicate that responsible evaluation and governance of LLMs require moving beyond scalar averages toward multidimensional, tail-sensitive risk profiling.

  • 3 authors
·
Jan 28 2

EvalGIM: A Library for Evaluating Generative Image Models

As the use of text-to-image generative models increases, so does the adoption of automatic benchmarking methods used in their evaluation. However, while metrics and datasets abound, there are few unified benchmarking libraries that provide a framework for performing evaluations across many datasets and metrics. Furthermore, the rapid introduction of increasingly robust benchmarking methods requires that evaluation libraries remain flexible to new datasets and metrics. Finally, there remains a gap in synthesizing evaluations in order to deliver actionable takeaways about model performance. To enable unified, flexible, and actionable evaluations, we introduce EvalGIM (pronounced ''EvalGym''), a library for evaluating generative image models. EvalGIM contains broad support for datasets and metrics used to measure quality, diversity, and consistency of text-to-image generative models. In addition, EvalGIM is designed with flexibility for user customization as a top priority and contains a structure that allows plug-and-play additions of new datasets and metrics. To enable actionable evaluation insights, we introduce ''Evaluation Exercises'' that highlight takeaways for specific evaluation questions. The Evaluation Exercises contain easy-to-use and reproducible implementations of two state-of-the-art evaluation methods of text-to-image generative models: consistency-diversity-realism Pareto Fronts and disaggregated measurements of performance disparities across groups. EvalGIM also contains Evaluation Exercises that introduce two new analysis methods for text-to-image generative models: robustness analyses of model rankings and balanced evaluations across different prompt styles. We encourage text-to-image model exploration with EvalGIM and invite contributions at https://github.com/facebookresearch/EvalGIM/.

  • 17 authors
·
Dec 17, 2024

Mediocrity is the key for LLM as a Judge Anchor Selection

The ``LLM-as-a-judge'' paradigm has become a standard method for evaluating open-ended generation. To address the quadratic scalability costs of pairwise comparisons, popular benchmarks like Arena-Hard and AlpacaEval compare all models against a single anchor. However, despite its widespread use, the impact of anchor selection on the reliability of the results remains largely unexplored. In this work, we systematically investigate the effect of anchor selection by evaluating 22 different anchors on the Arena-Hard-v2.0 dataset. We find that the choice of anchor is critical: a poor anchor can dramatically reduce correlation with human rankings. We identify that common anchor choices (best-performing and worst-performing models) make poor anchors. Because these extreme anchors are consistently better or worse than all other models, they are seldom indicative of the relative ranking of the models. We further quantify the effect size of anchor selection, showing it is comparable to the selection of a judge model. We conclude with actionable recommendations. First, we conduct a power analysis, and compute sufficient benchmark sizes for anchor-based evaluation, finding that standard benchmark sizes are insufficient for pairwise evaluation and fail to distinguish between competitive models reliably. Second, we provide guidelines for selecting informative anchors to ensure reliable and efficient evaluation practices.

  • 4 authors
·
Mar 17

Who Evaluates AI's Social Impacts? Mapping Coverage and Gaps in First and Third Party Evaluations

Foundation models are increasingly central to high-stakes AI systems, and governance frameworks now depend on evaluations to assess their risks and capabilities. Although general capability evaluations are widespread, social impact assessments covering bias, fairness, privacy, environmental costs, and labor practices remain uneven across the AI ecosystem. To characterize this landscape, we conduct the first comprehensive analysis of both first-party and third-party social impact evaluation reporting across a wide range of model developers. Our study examines 186 first-party release reports and 183 post-release evaluation sources, and complements this quantitative analysis with interviews of model developers. We find a clear division of evaluation labor: first-party reporting is sparse, often superficial, and has declined over time in key areas such as environmental impact and bias, while third-party evaluators including academic researchers, nonprofits, and independent organizations provide broader and more rigorous coverage of bias, harmful content, and performance disparities. However, this complementarity has limits. Only model developers can authoritatively report on data provenance, content moderation labor, financial costs, and training infrastructure, yet interviews reveal that these disclosures are often deprioritized unless tied to product adoption or regulatory compliance. Our findings indicate that current evaluation practices leave major gaps in assessing AI's societal impacts, highlighting the urgent need for policies that promote developer transparency, strengthen independent evaluation ecosystems, and create shared infrastructure to aggregate and compare third-party evaluations in a consistent and accessible way.

  • 35 authors
·
Nov 6, 2025

How Can Rhetoric Reward-Hack AI Reviewers? Dissecting Rhetorical Sensitivity in AI-Based Peer Review

As large language models increasingly participate in scientific evaluation, we investigate a potential form of reward hacking: how rhetorical choices shape AI-review judgments when reported scientific content is preserved and how these effects vary across evaluation conditions. We construct a controlled corpus of 4,200 full-paper manuscripts derived from 120 anonymized ICLR 2026 submissions. Two LLM rewriters transform six rhetorical dimensions in opposing directions, and five LLM reviewers evaluate the resulting manuscripts under standard and strict protocols. We also test joint, recursive, and reviewer-guided rewriting. Our results show that rhetorical sensitivity is structured rather than uniform. Evidence framing and novelty stance produce the largest positive-negative contrasts in overall assessment, with scope framing forming a weaker second tier; the remaining dimensions have smaller or less stable effects. This hierarchy persists across human-assessed quality levels, but score movement depends strongly on the AI reviewer's original score: lower scores tend to rise, higher scores tend to fall, and directional contrasts are clearest in the middle ranges. More elaborate workflows do not reliably yield larger gains. Joint rewriting is strongly rewriter-dependent, reviewer guidance does not consistently outperform an unguided second pass, and repeated rewriting yields diminishing, configuration-dependent returns. Across conditions, the rewriter primarily determines the separation between opposing variants, whereas the reviewer determines the magnitude and sign of their score effects. Strict review lowers mean OA by 1.36 points without consistently changing rhetorical sensitivity. These findings identify when rhetorical presentation influences AI scientific review and motivate evaluation systems robust to content-preserving variation in scientific writing.

Image-based Treatment Effect Heterogeneity

Randomized controlled trials (RCTs) are considered the gold standard for estimating the average treatment effect (ATE) of interventions. One use of RCTs is to study the causes of global poverty -- a subject explicitly cited in the 2019 Nobel Memorial Prize awarded to Duflo, Banerjee, and Kremer "for their experimental approach to alleviating global poverty." Because the ATE is a population summary, anti-poverty experiments often seek to unpack the effect variation around the ATE by conditioning (CATE) on tabular variables such as age and ethnicity that were measured during the RCT data collection. Although such variables are key to unpacking CATE, using only such variables may fail to capture historical, geographical, or neighborhood-specific contributors to effect variation, as tabular RCT data are often only observed near the time of the experiment. In global poverty research, when the location of the experiment units is approximately known, satellite imagery can provide a window into such factors important for understanding heterogeneity. However, there is no method that specifically enables applied researchers to analyze CATE from images. In this paper, using a deep probabilistic modeling framework, we develop such a method that estimates latent clusters of images by identifying images with similar treatment effects distributions. Our interpretable image CATE model also includes a sensitivity factor that quantifies the importance of image segments contributing to the effect cluster prediction. We compare the proposed methods against alternatives in simulation; also, we show how the model works in an actual RCT, estimating the effects of an anti-poverty intervention in northern Uganda and obtaining a posterior predictive distribution over effects for the rest of the country where no experimental data was collected. We make all models available in open-source software.

Don't Make Your LLM an Evaluation Benchmark Cheater

Large language models~(LLMs) have greatly advanced the frontiers of artificial intelligence, attaining remarkable improvement in model capacity. To assess the model performance, a typical approach is to construct evaluation benchmarks for measuring the ability level of LLMs in different aspects. Despite that a number of high-quality benchmarks have been released, the concerns about the appropriate use of these benchmarks and the fair comparison of different models are increasingly growing. Considering these concerns, in this paper, we discuss the potential risk and impact of inappropriately using evaluation benchmarks and misleadingly interpreting the evaluation results. Specially, we focus on a special issue that would lead to inappropriate evaluation, \ie benchmark leakage, referring that the data related to evaluation sets is occasionally used for model training. This phenomenon now becomes more common since pre-training data is often prepared ahead of model test. We conduct extensive experiments to study the effect of benchmark leverage, and find that it can dramatically boost the evaluation results, which would finally lead to an unreliable assessment of model performance. To improve the use of existing evaluation benchmarks, we finally present several guidelines for both LLM developers and benchmark maintainers. We hope this work can draw attention to appropriate training and evaluation of LLMs.

  • 9 authors
·
Nov 3, 2023

A Survey on Evaluation of Large Language Models

Large language models (LLMs) are gaining increasing popularity in both academia and industry, owing to their unprecedented performance in various applications. As LLMs continue to play a vital role in both research and daily use, their evaluation becomes increasingly critical, not only at the task level, but also at the society level for better understanding of their potential risks. Over the past years, significant efforts have been made to examine LLMs from various perspectives. This paper presents a comprehensive review of these evaluation methods for LLMs, focusing on three key dimensions: what to evaluate, where to evaluate, and how to evaluate. Firstly, we provide an overview from the perspective of evaluation tasks, encompassing general natural language processing tasks, reasoning, medical usage, ethics, educations, natural and social sciences, agent applications, and other areas. Secondly, we answer the `where' and `how' questions by diving into the evaluation methods and benchmarks, which serve as crucial components in assessing performance of LLMs. Then, we summarize the success and failure cases of LLMs in different tasks. Finally, we shed light on several future challenges that lie ahead in LLMs evaluation. Our aim is to offer invaluable insights to researchers in the realm of LLMs evaluation, thereby aiding the development of more proficient LLMs. Our key point is that evaluation should be treated as an essential discipline to better assist the development of LLMs. We consistently maintain the related open-source materials at: https://github.com/MLGroupJLU/LLM-eval-survey.

  • 16 authors
·
Jul 6, 2023 1

How Discriminative Are Your Qrels? How To Study the Statistical Significance of Document Adjudication Methods

Creating test collections for offline retrieval evaluation requires human effort to judge documents' relevance. This expensive activity motivated much work in developing methods for constructing benchmarks with fewer assessment costs. In this respect, adjudication methods actively decide both which documents and the order in which experts review them, in order to better exploit the assessment budget or to lower it. Researchers evaluate the quality of those methods by measuring the correlation between the known gold ranking of systems under the full collection and the observed ranking of systems under the lower-cost one. This traditional analysis ignores whether and how the low-cost judgements impact on the statistically significant differences among systems with respect to the full collection. We fill this void by proposing a novel methodology to evaluate how the low-cost adjudication methods preserve the pairwise significant differences between systems as the full collection. In other terms, while traditional approaches look for stability in answering the question "is system A better than system B?", our proposed approach looks for stability in answering the question "is system A significantly better than system B?", which is the ultimate questions researchers need to answer to guarantee the generalisability of their results. Among other results, we found that the best methods in terms of ranking of systems correlation do not always match those preserving statistical significance.

  • 3 authors
·
Aug 18, 2023

Evaluation Gaps in Machine Learning Practice

Forming a reliable judgement of a machine learning (ML) model's appropriateness for an application ecosystem is critical for its responsible use, and requires considering a broad range of factors including harms, benefits, and responsibilities. In practice, however, evaluations of ML models frequently focus on only a narrow range of decontextualized predictive behaviours. We examine the evaluation gaps between the idealized breadth of evaluation concerns and the observed narrow focus of actual evaluations. Through an empirical study of papers from recent high-profile conferences in the Computer Vision and Natural Language Processing communities, we demonstrate a general focus on a handful of evaluation methods. By considering the metrics and test data distributions used in these methods, we draw attention to which properties of models are centered in the field, revealing the properties that are frequently neglected or sidelined during evaluation. By studying these properties, we demonstrate the machine learning discipline's implicit assumption of a range of commitments which have normative impacts; these include commitments to consequentialism, abstractability from context, the quantifiability of impacts, the limited role of model inputs in evaluation, and the equivalence of different failure modes. Shedding light on these assumptions enables us to question their appropriateness for ML system contexts, pointing the way towards more contextualized evaluation methodologies for robustly examining the trustworthiness of ML models

  • 5 authors
·
May 10, 2022

Decomposing and Measuring Evaluation Awareness

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

  • 6 authors
·
May 20

TenetBench: A Clause-Anchored Russian Corpus for Evaluating Language Models against Professional Codes of Ethics

Codes of professional ethics guide and restrict specialists, who increasingly work with assistants built on large language models. Such an assistant can breach the user's code, or help the user breach it, without the user noticing. Evaluations reporting broad categories of harm cannot show how often this happens or which provision is breached. TenetBench pairs 3,433 requests conflicting with clauses of fifteen Russian-language professional codes with benign twins that differ only in the legitimacy of what is asked. An automated pipeline generates the tasks from the text of a code, so the corpus grows with every code added and a provision yields new tasks on every run. It also produced tasks and human-validated verdicts from two English codes of other jurisdictions. In human validation of the frozen pipeline, specialists rated 0.995 of the conflicting tasks plausible, and the model judge agreed with the adjudicated label at Gwet's AC1 of at least 0.894 on 1,860 elements. We evaluated fourteen model and reasoning-mode combinations on whether they produce the conflicting text, decline or warn. Eleven of the fourteen produce conflicting content on most such requests, mostly without a warning, and nearly three warnings in four accompany conflicting content. The models that conform more do so by declining while still serving the legitimate twin, and enabling reasoning halved the refusals of two models. We hope the dataset and pipeline will help evaluate how language models observe professional norms and let diagnostic datasets be built for particular professions and organisations.

  • 5 authors
·
Oct 20 6