LLM Evaluation

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

194 papers in the last four weeks, up 87% on the four weeks before. 1.9% of all new papers.

Jul 13Week of Sep 28

Latest papers 1,643

Jul 20, 2026cs.CL

VEHBench: A Stage-Local Diagnostic Benchmark for LLM-Assisted Vibration Energy Harvester Design

Battery-free Internet of Things (IoT) requires iterative design of vibration energy harvesters (VEHs) under coupled physical constraints, while LLMs are emerging as interface layers for engineering workflows. However, existing engineering benchmarks primarily assess final artifact validity, offering limited insights into how LLMs behave across different stages of coupled physical design. We introduce VEHBench, an engineering-native diagnostic benchmark for LLM-assisted VEH design, featuring 763 literature-grounded tasks scored by an analytical physical oracle. VEHBench evaluates four design roles: specification triage, verifier-guided search, corrupted-state recovery, and policy-conditioned selection. Experimental results reveal that LLM capability is strongly stage-dependent: no single model consistently dominates the entire workflow, and response-control profiles expose distinct behavioral patterns across design roles. VEHBench thus provides a stage-aware foundation for evaluating, selecting, routing, and improving verifier-grounded engineering LLMs. The benchmark artifact is available at https://huggingface.co/datasets/AnonymousVehbench/vehbench

Depeng Su, Yuyu Luo, Guobiao Hu

1Internet of Things Thrust, Information Hub 2Data Science and Analytics Thrust, Information Hub The Hong Kong University of Science and Technology (Guangzhou)

Jul 20, 2026cs.LG

Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints

Structure-based drug design (SBDD) leverages the 3D structure of protein targets, often complemented by other spatial constraints, to generate candidate binding molecules. While diffusion models have dominated as a leading paradigm for high-quality 3D molecule generation, LLM-based methods are rapidly emerging in molecular design and have shown competitive performance in pocket-conditioned molecular generation. However, their ability to reason about physics and 3D spatial environments is largely underexplored. In this work, we systematically analyze whether current general-purpose LLMs are capable of navigating complex 3D constraints compared to established baselines such as specialized diffusion models. We consider 3D ligand generation conditioned on protein pockets together with ligand- and interaction-derived spatial constraints, including anchor fragments, pharmacophore points, and mandatory pocket-ligand interactions. To enable this evaluation, we introduce 3D-Fit - a token-efficient benchmarking strategy for assessing LLM performance on multi-conditioned spatial molecule generation. Our findings reveal a clear pattern in LLM spatial capabilities: while they still lag behind state-of-the-art approaches, they are promising and can handle multiple spatial constraints simultaneously, enabling scaling to heterogeneous setups.

Thomas MacDougall, Maksim Kuznetsov, Roman Schutski +5

Insilico Medicine Canada Inc. · Insilico Medicine Hong Kong Ltd. · Insilico Medicine AI Limited

Constrained Generative Modeling LLM Evaluation Structure-Based Drug Design 3D Molecular Generation Molecular Generation Protein Binder Design

Jul 20, 2026cs.AI

WorldCupArena: Fine-Grained Evaluation of Language Models and Deep-Research Agents on Football Forecasting

Predicting a football match before kickoff requires more than knowing past results: a model must use changing information and make a clear prediction before the answer is available. We present WorldCupArena, a dynamic benchmark for language models and deep-research agents. The 2026 FIFA World Cup is its first evaluation, and the same process can be reused for future leagues and cups. Before each match, a model either receives a common evidence package or searches for information itself. It predicts the result and score, likely players and events, match statistics, and the outcome of the competition. After the match, these predictions are compared with the recorded result. We report result accuracy, exact-score accuracy, and a scoreline score that gives some credit when a predicted score is close but not exact, together with scores for the other prediction tasks. Across systems, similar result accuracy can mask larger differences in detailed predictions. Four systems predicted champion Spain, and two of them also recovered the exact final pairing. Compared with betting-market and human-fan baselines, the best system shows only small gains in result and exact-score accuracy, but a clearer gain in Scoreline. New schedules can be added as they begin, allowing the benchmark to evaluate future models without using outcomes that are already known. Code, predictions and evaluation scripts will be publicly released.

Zhaokai Wang, Tianlin Gui, Jiayuan Rao +3

Shanghai Jiao Tong University · Shanghai, China · Nanjing University +4

LLM Evaluation Forecasting Benchmarks Probabilistic Forecasting Sports Outcome Prediction Agent Benchmarks

Jul 20, 2026cs.NI

Human Grounded Evaluation of Large Language Models for Optical Network Automation

Large language models (LLMs) are increasingly adopted for network automation, yet their output quality and inference cost can vary substantially across LLM families. We present HuGLEN, a stepwise evaluation pipeline that uses an LLM-as-a-judge together with a small set of expert ratings to enable scalable and reproducible comparison of candidate LLMs, and to rank them using a quality efficiency score (QES). We demonstrate HuGLEN for translating outputs from an explainable artificial intelligence (XAI) model for the optical network quality of transmission (QoT) estimation task into operator-friendly explanations. Our results show that a medium-sized LLM (12B parameters) achieves the highest QES, indicating the best trade-off between explanation quality and efficiency. Overall, HuGLEN reduces the human-labeling burden while supporting consistent model selection for operator-facing automation tasks.

Kiarash Rezaei, Omran Ayoub, Paolo Monti +1

Department of Electrical Engineering, Chalmers University of Technology, 412 96 Gothenburg, Sweden · University of Applied Sciences and Arts of Southern Switzerland, 6928 Lugano, Switzerland

LLM Inference Efficiency LLM Evaluation Human-in-the-Loop Evaluation Automated Evaluation

Jul 20, 2026cs.CL

Pancasila-Dilemmas: Evaluating Large Language Models on Indonesian Human Value Dilemmas Grounded in Pancasila

The value alignment of large language models (LLMs) is crucial for ensuring responses align with human intention and value preferences. However, most evaluations of value alignment focus on Western or universal values, while assessments grounded in the value systems of specific countries remain scarce. In this paper, we introduce Pancasila-Dilemmas, an evaluation dataset of 1,834 questions derived from Indonesian news, classified by 5 values of Pancasila: Religion, Humanity, Unity, Democracy, and Social Justice. This dataset reflects daily life in Indonesia, making it suitable for measuring the value alignment of LLMs deployed for Indonesia. To ensure a more rigorous evaluation, we choose scenarios containing dilemmas. The dataset is proofread by native speakers and answered by 5 diverse Indonesian citizens. We evaluate 50 closed- and open-source LLMs on our dataset. Results reveal that all evaluated LLMs achieves less than 0.5 Probability Match Score (PMS) and 0.72 Max-Vote Agreement Score (MVAS). Compared by each values, LLMs mostly struggle in Religion and Unity dilemma cases. This highlights a significant gap in capturing Indonesian values. The dataset is publicly available at https://github.com/tjunlp-lab/Pancasila-Dilemmas.

Supryadi, Irfan, Julianti +4

TJUNLP Lab, School of Computer Science and Technology, Tianjin University, China · Universitas Universal, Indonesia · Beijing Language and Culture University, Beijing, China +2

LLM Evaluation LLM Alignment Moral Reasoning in Language Models Cultural Alignment Value Alignment

Jul 20, 2026cs.CL

Evaluating Federated Pre-Training: On the Reliability of Downstream Fine-Tuning and Intrinsic Evaluation

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

Claudia Grosser, Maike Heuer, Denis Krompass +1

Technical University Munich · Siemens AG

Language Model Pretraining LLM Evaluation Federated Learning

Jul 20, 2026cs.CR

HALLMARK: Diagnosing Three Failure Modes in LLM Citation Verifiers

Large language models (LLMs) now routinely draft literature reviews and assist with academic writing, which means a higher risk of fabricated references: GPTZero found 53 papers with hallucinated citations among NeurIPS 2025's accepted set. Rule- and LLM-based verifiers are emerging, but no shared benchmark compares them and gives detailed failure diagnostics. We close that gap with HALLMARK (Hallucination benchmark): 2,526 BibTeX entries spanning 14 hallucination types, three difficulty tiers, six diagnostic sub-tests per entry, and a contamination-resistant held-out split. On it we evaluate a DOI-lookup baseline, frontier LLMs zero-shot, tool-augmented agents, and our own rule-based, co-designed verifier bibtex-updater. Across the benchmark one result is consistent: the false-positive rate, not recall, decides whether a verifier is deployable. HALLMARK makes it concrete through three failure modes: agentic lookups buy recall but inflate false positives; at a venue-realistic base rate, the order-of-magnitude spread in false-positive rates (FPRs) -- not recall -- governs whether a verifier's flags are mostly true catches or mostly noise; and most LLMs over-flag papers published past their training cutoff, where only the two latest-cutoff models hold their false-positive rate near in-distribution levels (a signal we report as descriptive, since it is confounded with possible recall of those entries). Thus FPR is the deployment bottleneck, but an undetected fabrication remains the costlier error for the scientific record.

Patrik Reizinger, Wieland Brendel

Max Planck Institute for Intelligent Systems, ELLIS Institute Tübingen

LLM Evaluation Citation Verification LLM Hallucination Detection

Jul 20, 2026cs.DL

Benchmarking Resource-Efficient LLMs for Research Topic Ontology Generation in the Biomedical Field

Knowledge Organization Systems like Ontologies and taxonomies are fundamental for structuring scientific knowledge, yet their manual curation presents a persistent bottleneck in knowledge management. While Large Language Models (LLMs) offer a scalable mechanism for automated ontology generation, their capacity to classify complex, domain-specific semantics requires systematic evaluation. In this paper, we assess the performance of five small, open-source LLMs (up to 9 billion parameters) in identifying semantic relationships between biomedical concepts. To support this evaluation, we introduce MeSH-Rel-4K, a dataset comprising 4K semantic relationships extracted from the Medical Subject Headings (MeSH). We analyse three adaptation strategies: standard prompting, Chain-of-Thought prompting, and fine-tuning. While parameter-constrained models traditionally struggle with the nuances of in-context logic, our results reveal that targeted fine-tuning increases the average F1-score by 34.1 percentage points. These results confirm that direct fine-tuning effectively exceeds the reasoning bottlenecks of smaller LLMs, providing an accurate, automated methodology for the construction and evolution of specialised biomedical ontologies.

Tanay Aggarwal, Angelo Salatino, Francesco Osborne +1

Knowledge Media Institute, The Open University, Milton Keynes, UK · Department of Business and Law, University of Milano-Bicocca, Milan, IT

Relation Extraction LLM Evaluation LLM Prompting LLM Fine-Tuning Medical LLMs Ontology Learning

Jul 20, 2026cs.AI

WuYu-EnvLE-Bench: A Benchmark for Evaluating Large Language Models in Environmental Law Enforcement

Large language models (LLMs) are increasingly considered for environmental enforcement, but their ability to produce traceable enforcement decisions remains unclear. We introduce WuYu-EnvLE-Bench, a benchmark built from real enforcement cases, regulatory standards, and expert review. It contains 2,521 benchmark instances, 14 tasks, and 12 pollution-medium subdomains across pre-enforcement, in-enforcement, and post-enforcement workflows. Using Absolute Environmental Enforcement Score (AES) and Intelligent Enforcement Index (IEI), we evaluate open-source and closed-source LLMs across capability, response quality, and resource efficiency. Results show that LLMs perform well on rule-bounded tasks but remain unreliable in evidence-chain construction, contradiction detection, multi-source integration, and procedural judgment. Model scaling also shows diminishing returns: medium-sized models approach leading models in structured tasks, while larger models do not reliably overcome evidence-reasoning bottlenecks. WuYu-EnvLE-Bench highlights the need for evidence-grounded, rule-aware, and task-adaptive enforcement reasoning.

Ziliang Yang, Yi Zhang, Kaijun Lin +3

School of Environment, Tsinghua University · College of Economics and Management, Beijing University of Technology · State Key Laboratory of Iron and Steel Industry Environmental Protection, School of Environment, Tsinghua University +1

LLM Evaluation LLM Grounding Legal Reasoning

Jul 20, 2026cs.CL

Large Language Models for Citation Function Classification

Citation function classification plays a crucial role in understanding the relationships between scientific publications and advancing bibliometric analysis. This study presents one of the first comprehensive evaluations of multiple state-of-the-art (SOTA) large language models (LLMs) for citation function classification, achieving new SOTA results on the ACL-ARC dataset. We systematically compare five models (Mistral 7B, Orca 2-7B, LLaMA 3.1-8B, Falcon 7B, and SciBERT) across zero-shot, few-shot, and fine-tuning approaches. Our fine-tuned Falcon 7B model achieves a 73.3% macro F1 score on ACL-ARC, representing a significant improvement over previous methods. Additionally, we introduce AC3, a novel dataset featuring a seven-category annotation scheme that distinguishes between neutral acknowledgments and explicit evaluative stances (more opinion-oriented citations - criticizing, complimenting, contradicting). The dataset is implemented across four context extraction variants to systematically evaluate the impact of contextual scope on classification performance. We also provide detailed analysis of model performance, experimental configurations, and limitations to guide future research in this domain. To our knowledge, this is one of the first studies dedicated to comprehensive model comparison for citation function classification, addressing a gap identified in recent surveys.

Daniel Vodička, Jakub Šmíd, Pavel Král +1

Department of Computer Science and Engineering · NTIS – New Technologies for the Information Society University of West Bohemia in Pilsen, Faculty of Applied Sciences Univerzitní 2732/8, 301 00 Pilsen, Czech Republic · CNRS LORIA, Université de Lorraine, Nancy, France

LLM Evaluation LLM Fine-Tuning Text Classification Bibliometric Analysis

Jul 20, 2026cs.CL

Oracle Gap and Signal Fidelity: A Fixed-Pool Diagnostic for Test-Time Collaboration

Test-time collaboration, including self-consistency, best-of-N selection, critic models, and verifier pipelines, is often credited with broadly improving LLM reasoning, yet its gains are uneven and sometimes negative. We ask when training-free collaboration should be expected to help. For a fixed candidate pool, we decompose a selector or verifier's net gain into measurable factors: recoverable mass, verification-signal coverage, conditional selection quality, and harm to already-correct outputs. This reframes collaboration as a candidate-selection problem rather than as an intrinsic property of a multi-agent topology. Across LiveCodeBench, MATH Level-5 hard subjects, and GPQA-Diamond, gains are bounded first by the oracle gap and then by signal fidelity, which we measure directly as candidate-level agreement between verifier verdicts and official labels. On LiveCodeBench, a public-test verifier (MCC 0.825) gains +8.14 percentage points (pp) over a first-sample baseline; a generated-test verifier (MCC 0.248) improves by +2.70pp and is not statistically distinguishable from an LLM selector, but operates at near-zero harm versus the selector's 4.69% harm rate. On MATH, a symbolic answer-equivalence selector beats self-consistency by +4.67pp, while LLM selectors are negative. On GPQA-Diamond, recoverable mass is only 3.03% and 87.54% of candidate pools are answer-identical; a weaker model's pools shrink both further, suggesting that oracle gap is a joint property of task, model, and sampling configuration. Our framework yields a practical pre-deployment diagnostic: estimate the oracle gap, then measure coverage, signal fidelity, and harm before investing in collaboration.

Jie Hu

Research Institute of China Telecom

LLM Evaluation LLM Answer Verification Test-Time Scaling Multi-LLM Collaboration

Jul 19, 2026stat.ML

Efficient Sequential Evaluation of Large Language Models

We study the problem of sequentially evaluating a new large language model (LLM) on a fixed question set using historical performance data from prior LLMs. Our goal is to construct a confidence sequence (CS) for the model's capability on this question set and to design active querying rules that shrink the CS width as quickly as possible. For CS construction, we invert a family of test supermartingales and focus on two representative approaches: a reverse information projection (RIPr)-based approach and a testing-by-betting-based approach. We first study these approaches under an oracle setting, and demonstrate the oracle optimality of the RIPr-based construction. We then propose a growth-oriented querying rule that aims to maximize the worst-case one-step expected log-increment over the endpoints of the current CS. In practice, we build these test supermartingales and the querying rule on predictions of question-level correctness learned from historical data. We then analyze the shrinkage behavior of the resulting CSs and identify two key factors that slow the shrinkage rate of CSs: accumulated prediction mismatch and the spikiness of the querying distribution. Finally, motivated by this analysis, we propose several mixture querying rules that combine growth-oriented querying, prediction refinement, and uniform exploration, trying to mitigate the effects that slow the shrinkage rate. We provide experiments comparing different querying rules for the RIPr-based and testing-by-betting-based CSs across several synthetic testing datasets. Interestingly, we observe that the simplest querying rule, uniform sampling, can sometimes outperform more adaptive querying rules for both methods.

Chia-Yu Hsu, Shubhanshu Shekhar

EECS Department, University of Michigan

LLM Evaluation Confidence Estimation in Language Models Long-Context Language Model Evaluation Active Learning

Jul 19, 2026cs.CL

Literary Non-Style in LLM-Generated Text

Prior work on LLM-generated text has demonstrated quantitative and qualitative departures from text produced by humans. LLM-generated texts differ from human writing in style, resulting in a characteristic textual "feel," while the semantic range of LLMs is much restricted compared to that of humans. In this contribution, I note simple but consistent patterns in the statistical distribution of n-grams within LLM-generated text. Via qualitative analysis of these n-grams, I reveal deficiencies in LLM style. Because higher-order n-grams correlate to semantic content, I conclude that questions of style and semantics are not cleanly separable.

Cory Massaro

Jul 19, 2026cs.AI

A Systematic Evaluation of Trajectory Data Curation for LoRA Fine-Tuning of Code Agents

Supervised fine-tuning (SFT) of open-weight LLMs on expert agent trajectories has emerged as a prominent approach to building capable code agents without reliance on proprietary models. A central yet underexplored question is how trajectory quality and quantity jointly shape model performance. We present a systematic empirical study of trajectory data filtering for LoRA fine-tuning of Qwen2.5-Coder-7B-Instruct on the SWE-trajectory dataset (67,074 trajectories, of which 32,161 are resolved). We propose a two-axis quality scoring framework -- Efficiency and Style -- and evaluate it through 16 controlled experiments spanning strategy, scale, and ablation analyses. Since 7B-scale models attain near-zero SWE-bench resolve rates, we adopt cross-entropy (CE) loss on held-out trajectories as the primary metric, validated via first-action generation: CE loss and ROUGE-L are perfectly rank-correlated (Spearman

ρ

= -1.00), with limited-sample evidence supporting but not conclusively establishing this proxy. Our results reveal a scale-dependent quality-quantity trade-off: at small scales, doubling the dataset (500 to 1,000) yields ~12.7% CE-loss reduction whereas the TopQ-Random gap stays <1% (Mann-Whitney p > 0.10); at 2,000 trajectories this same gap widens to 3.6% (p = 0.016). Ablation further identifies error-retry rate as the dominant sub-dimension, performing comparably to the full composite (

Δ

< 0.2%). Together, these findings establish trajectory-level quality scoring as a viable but scale-sensitive lever for code-agent SFT and offer a proxy-validated evaluation protocol for the regime where end-to-end resolve rate is statistically infeasible.

Yunze Han

Vdynamics GmbH, 80935 Munich, Germany

Supervised Fine-Tuning LLM Evaluation AI Coding Agents

Jul 19, 2026cs.CL

A Consensus-Based Framework for Relative Preference Evaluation of Large Language Models

Traditional benchmarks for LLMs primarily rely on static datasets and objective scoring metrics, which often fail to capture differences in response quality when multiple answers are acceptable. In such settings, correctness alone is insufficient to distinguish between responses that vary in clarity, completeness, and usefulness. This paper introduces a consensus-based evaluation framework that measures relative preference among model-generated responses rather than absolute correctness. Instead of evaluating outputs against a fixed ground truth, we assess how a panel of diverse LLMs ranks anonymized candidate responses to the same prompt. This approach treats aggregate inter-model agreement as a proxy for perceived response quality under blind conditions. We conduct a controlled study using five state-of-the-art LLMs across multiple domains, including programming, general knowledge, safety, logical reasoning, and mathematics. Each model generates responses and independently ranks peer outputs through a structured voting process. Scores are aggregated into a Relative Intelligence Index (RII), representing how frequently a model's responses are preferred by other models. Our findings reveal consistent preference patterns across domains, with certain models more frequently ranked highly by their peers. However, we emphasize that these results reflect inter-model preference alignment rather than objective correctness or human judgment. This framework provides a scalable, model-driven method for comparative evaluation, offering an alternative perspective on response quality in scenarios where multiple valid answers exist. While not directly aligned with human evaluation, prior work suggests that aggregated model preferences can partially correlate with human judgments, motivating this as a proxy signal.

Mohtashim Khan

LLM Evaluation LLM-as-a-Judge

Jul 19, 2026cs.CR

A Systematic Evaluation of Traditional Privacy Policy Analysis Tools Against LLMs

The advent of LLMs has significantly changed the research on privacy policy and data compliance analysis by enabling tasks that previously required specialized, domain-specific tools. However, it remains unclear to what extent LLMs can truly replicate the diverse functionalities, and the wide range of methodologies and analysis offered by prior work. In this paper, we conduct the first systematic evaluation of whether off-the-shelf LLMs can replace specialized privacy analysis tools. We study six representative tools spanning three major functionalities: contradiction detection, regulatory compliance analysis, and privacy policy summarization and aggregation, and across three intermediate tasks: structured data extraction using tuples, Semantic Role Labeling (SRL) and manual privacy policy labeling. We compare the performance of two state-of-the-art LLMs (GPT-5.2 and Gemini-2.5 in various configurations) against the tools by directly prompting the models to perform corresponding functionalities and tasks on a custom dataset of 10 privacy policies, allowing us to assess whether off-the-shelf models can produce tool-specific functionalities without further engineering or domain-specific training, major limitations in prior work. Our results show that LLMs consistently match or exceed the capabilities of existing tools across the functionalities. In manual labeling of first-party collection entities, LLMs achieved an average precision of 81.8% and recall of 70.9%, while for labeling of third-party sharing entities, they achieved an average precision of 91.4% and recall of 70.8% compared to the OPP-115 dataset. Overall, our findings indicate that LLMs can effectively perform a broad range of functionalities and tasks in privacy policy and regulation analysis that previously required specialized tools.

Madhav Aryal, Sudipa Saha, Sunil Manandhar +2

University of South Florida · IBM T.J. Watson Research Center

Privacy Auditing LLM Evaluation Text Summarization Legal Document Analysis

Jul 19, 2026cs.AI

When LLMs Over-Answer: Measuring and Mitigating Quality Issues in LLM-Based Hardware Description Language Question Answering

The rapid advancement of large language models (LLMs) has led practitioners to increasingly rely on them for answering questions about hardware description languages (HDLs). Because HDL is ultimately synthesized into physical hardware, an imprecise or redundant answer can propagate into timing violations or non-synthesizable logic that surface only late in the design flow, making the quality of HDL answers especially consequential. However, the quality of LLM-generated responses, particularly in comparison with answers provided by human experts, remains unclear. To investigate this question, we collect 6,246 HDL Q&A posts with accepted answers from Stack Overflow and curate them into a dataset, organized into a taxonomy of four main categories (Conceptual, Debugging, Generation, and Optimization) and ten subcategories. Using this dataset, we design a user study conducted with 19 HDL engineers with one to three years of experience. Our findings reveal a pervasive over answering tendency: LLMs supply correct content but bury it under redundant alternatives (65.7%) and verbose padding (69.1%), while nearly half of answers (49.0%) fail to fully align with expert answers yet participants still preferred LLM responses for readability (58.3%). Motivated by these findings, we propose a multi-agent framework for improving LLM-based HDL question answering. We evaluate answer quality using an LLM-as-Judge and two structural metrics: the number of core answers, which reflects redundancy since LLMs often provide multiple alternative solutions, and the length of non-core content, which reflects verbosity. Evaluated on the four mainstream LLMs, our framework increases the average core-answer quality score from 3.71 to 4.67 (+0.96) and the non-core content quality from 3.72 to 4.23 (+0.51), on a five-point scale.

Ziteng Hu, Jiachi Chen, Wenhao Lv +2

Micro-Electronics Research Institute, Hangzhou Dianzi University Hangzhou, Zhejiang, China · The State Key Laboratory of Blockchain and Data Security, Zhejiang University Hangzhou, Zhejiang, China · College of Information Science and Engineering, Hunan Normal University Changsha, Hunan, China

LLM Evaluation Electronic Design Automation

Jul 19, 2026cs.LG

Solver-Hard Is Not Model-Hard: A Hardness-Controlled Diagnostic for LLM Constraint Reasoning

LLM constraint reasoners are often evaluated near the random-SAT phase transition, confounding density and solver hardness. We test instance-level transfer while near-matching clause density. At aligned size bins, with near-matched density and matched maximum clause width, we compare proof-hard expander-Tseitin and proof-easy ladder-Tseitin formulas, pigeonhole anchors, and density-mismatched controls. Theory separates their resolution hardness; a solver-specific Glucose mean-conflict proxy differs by up to

51\times

, and five other solvers preserve the direction. Across three included models (243 instances each; a fourth is excluded for abstention), the near-matched-density accuracy gaps range from

-32

to

+20

points, with a pooled gap of

+1.7

points (

p=0.74

) and a wrong-signed correctness-versus-conflict association (

r=+0.15

). A proof-preserving relabeling lowers accuracy in all five clusters for one model (mean

-93

points) but not another, exposing model-surface sensitivity. In a preregistered extension, provider-reported completion-token spend does not consistently increase with the proxy after accounting for formula length and censoring. At 16k, the reasoning model spends more on proof-easy matched formulas and exhausts its budget on the solver-easiest UNSAT family; the 32k C1 gap is absent. These scoped dissociations concern verdict accuracy and observed token spend, not certificate solving, exact proof length, or allocation efficiency.

Lucky Verma

Independent Researcher

LLM Evaluation Constraint Satisfaction LLM Reasoning

Jul 18, 2026cs.CL

JOR-Bench: Japanese Operations Research Benchmarks for Large Language Models

We present JOR-Bench, a collection of five Japanese-language benchmarks for evaluating the ability of large language models (LLMs) to formulate and solve operations research (OR) problems. Each benchmark is a Japanese translation of an existing English benchmark: IndustryOR, MAMO Complex LP, NL4OPT, OptiBench, and OptMATH, covering 1,319 problems spanning linear programming, mixed-integer programming, non-linear programming, and combinatorial optimization. JOR-Bench is a solver-independent benchmark that can be used with any solver or programming language, and consists of pairs of Japanese problem statements and expected numerical answers. We evaluate seven LLMs, including multilingual general-purpose models and Japanese-specialized models, on both the original English and the new Japanese versions, and compare performance across languages. For the main evaluation, we standardize execution with the Python interface to OR-Tools to make model outputs comparable and reproducible with open-source software. Our results show that OR formulation ability is largely language-neutral for strong multilingual models; the overall average accuracy difference between English and Japanese is only

-0.3

pp. Yet error analysis reveals subtle cross-lingual differences, including a pragmatic disambiguation failure in some domains that causes models to output decision-variable values instead of the objective value when the prompt is in Japanese.

Yuu Jinnai

CyberAgent Tokyo, Japan

Multilingual Language Model Evaluation LLM Evaluation Natural Language to Optimization Modeling

Jul 17, 2026cs.CL

Encoding EEG Signals to Examine Human-Like Next-Word Prediction Behaviour in Language Models

Language models (LMs) are trained to excel at predicting the next word in the sequence given prior context, and humans also share this predictability in reading comprehension. Neuroscience research reveals that next-word predictability influences brain response, as recorded at millisecond resolution using electroencephalography (EEG). While our evidence indicates that advanced LMs achieve accuracies closely aligned with human performance at the next-word prediction task, this raises the question: Does higher prediction accuracy necessarily mean that these models adequately capture the cognitive signals associated with human reading comprehension? Here, we generate regressors for both humans and LMs based on two information measures, including top-1 prediction and surprisal, to predict event-related potential (ERP) elicited from EEG recordings which reflect different stages of cognitive processing during reading. We argue that modelling ERP patterns offers fine-grained analysis of the cognitive plausibility of various LMs during reading. Our results indicate that only surprisal potentially correlates with language-processing ERPs, especially for open-class words with high semantic content. Moreover, our findings challenge the assumption that scaling LMs with increased parameters and computational budgets will consistently lead to improved convergence with human-like linguistic processing.

Boi Mai Quach, Binh T. Nguyen, Cathal Gurrin +1

School of Computing, ML-Labs, Department of Computer Science, School of Computing, ADAPT Centre, Dublin City University, Ireland · VNUHCM – University of Science · School of Computing, ADAPT Centre, Dublin City University, Ireland

Jul 17, 2026cs.AI

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.

Vipul Gupta, Zihao Wang, Razvan-Gabriel Dumitru +3

Scale AI

LLM Evaluation LLM Fine-Tuning

Jul 17, 2026cs.LG

StabilityBench: Benchmarking Instability in LLMs

AI Assistants are increasingly deployed in high-stakes settings, such as healthcare or government services. Yet their real-world behavior remains poorly understood due to strong context dependence. Current evaluation protocols follow a defense-in-depth paradigm with compounding layers of safeguards, ranging from traditional benchmarks to live or adversarial testing. Such benchmarks remain largely static and single-turn, limiting their ability to capture real-world variability in conversational settings. We propose StabilityBench, a principled, general and model-agnostic benchmark operator that transforms single-turn benchmark queries into multi-turn interaction histories. StabilityBench augments existing benchmarks by injecting realistic user simulations, through demographic proxies or sycophantic baits, while preserving original task intent. We apply StabilityBench to four benchmarks spanning mathematical reasoning, health question-answering and safety, and evaluate nine large language models under these conditions. Our results show that model performance is consistently unstable under these injections, with considerable performance degradations on three out of four benchmarks studied. These highlight important limitations of static evaluations and motivate more realistic evaluation settings. To this end, we propose StabilityBench-Mini: a size-preserving variant of StabilityBench that samples across diversification axes, enabling more realistic evaluation without increasing costs.

Emma Kondrup, Zachary Yang, Anne Imouza +1

1Mila — Quebec AI Institute · 2McGill University · 3Ubisoft La Forge

LLM Evaluation Language Model Generation Evaluation Multi-Turn Dialogue Evaluation Language Model Robustness

Jul 17, 2026cs.CL

Frontier AI performance across the business disciplines: a case-grounded benchmark of knowledge work and analytical reasoning

Large language models (LLMs) are improving rapidly as reflected in benchmark scores, yet these AI benchmarks largely test capabilities such as factual recall, narrow question answering, mathematical problem-solving, and coding and agentic tool-use. What remains poorly measured is AI progress on the analytical knowledge work white-collar professionals perform daily, including synthesizing complex information, exercising judgment under uncertainty and incomplete information, applying strategic and adversarial thinking in multi-stakeholder settings, weighing trade-offs, and producing defensible, structured analyses. This gap is even more pronounced for subjective components of such work, where success can be challenging to define. The "case method" form of education practiced by top business schools provides a natural foundation for addressing this measurement gap, and we construct BusinessCaseBench, a benchmark spanning hundreds of questions drawn from business cases across eighteen disciplines, each paired with a grading rubric derived from the expert-written instructor case solution. On BusinessCaseBench, frontier AI models already score highly against instructor rubrics, and capability within one model family improves substantially over two years. These results provide strong evidence that AI performance on this class of work is already high and rapidly improving, with implications for business schools, where case pedagogy trains undergraduates and MBAs in this kind of analytical reasoning, and for entry-level professional roles, where such skills have historically anchored early-career work.

Ajay Patel, Kartik Hosanagar, Ramayya Krishnan +2

Operations, Information and Decisions Department, The Wharton School, University of Pennsylvania, Philadelphia, 19104, PA, USA · Heinz College of Information Systems and Public Policy, Carnegie Mellon University, Pittsburgh, 15213, PA, USA · Department of Computer and Information Science, University of Pennsylvania, Philadelphia, 19104, PA, USA +1

LLM Evaluation Benchmark Design

Jul 17, 2026cs.CL

Before the Action: Benchmarking LLMs on Prospective Hypothesis Discovery

Large language models (LLMs) excel at answering pre-specified questions, yet their ability to navigate the open-ended, pre-conclusion stage of discovery remains largely unmeasured. We introduce Prospective Hypothesis Discovery (PHD), which asks models to autonomously construct grounded, discriminative, and testable hypothesis spaces from inconclusive evidence, including anomalous observations and fragmented records, to guide subsequent investigation. To evaluate this capability, we introduce HypoArena, comprising HypoData, a benchmark of 988 cases across six scientific and analytical domains, and HypoEval, an evaluation framework for open-ended hypothesis sets. To construct HypoData at scale, we propose Retrospective Context Regression, a Forge--Audit pipeline that reconstructs pre-conclusion contexts from completed expert documents by removing explicit conclusions, target hypotheses, and retrospective causal attributions while preserving the factual substrate. Because PHD admits multiple valid outputs, HypoEval combines bidirectional pairwise judgments with Bradley--Terry--Davidson aggregation for ranking and six-dimensional rubric scoring for diagnosis. Experiments on 15 frontier LLMs reveal clear capability stratification and model-dependent effects of structured analytical skills, with gains for several lower-performing models on HypoArena but regressions for other systems, including a top-performing model. Compared with absolute rubric scoring, arena evaluation resolves finer-grained differences among models, with aggregated rankings showing strong agreement with human experts and an independent judge. Together, these results support treating PHD as a distinct target for evaluating how LLMs formulate investigative directions when final conclusions are withheld. Our code and data are publicly available at github.com/SKYLENAGE-AI/HypoArena and github.com/SKYLENAGE-AI/HypoArena.

Tianyun Zhong, Wangyi Jiang, Wei Wang +15

University of Chinese Academy of Sciences · Institute of Software, Chinese Academy of Sciences · 3Alibaba Group

LLM Evaluation Scientific Hypothesis Generation Large Language Model-Guided Scientific Discovery

Jul 17, 2026cs.AI

Neuro-Symbolic AI for LEED compliance: Document-Centric Benchmarking, Deterministic Numeric Checking, and When Multimodal Hurts

LEED v4.1 BD+C certification remains a document-intensive process that requires reviewers to read hundreds of pages of project evidence and apply credit-specific threshold logic by hand. This paper investigates whether small, locally deployed language models can perform meaningful screening of LEED documentation and how deterministic symbolic components should share that work. A neuro-symbolic pipeline is introduced that aligns project PDFs to LEED credit sections, retrieves evidence with credit-aware keyword signatures, verifies compliance with a locally hosted 4-billion-parameter language model, and applies a LEED-specific numeric checker to quantitative thresholds. Experiments on four university buildings (484 PDFs, 153 credit-level decisions) show that a 4-billion-parameter model (gemma3:4b) is the strongest text-only core verifier, achieving 67.3% accuracy and outperforming a larger 8-billion-parameter model (llama3.1:8b) in this task. The deterministic numeric checker corrects arithmetic errors on key quantitative credits, moving EA-p2 from 50% to 100% accuracy and improving several other credits when required values are reliably extracted. At the same time, the full neuro-symbolic configuration achieves 61.6% overall accuracy, trailing the best text-only baseline due to extraction failures and conservative behavior on qualitative categories. Systematic ablations show that adding low-resolution drawing images (150-300 dpi) consistently reduces accuracy, and that prompt effectiveness depends on the building's ground-truth PASS rate: rubric prompts perform best on documentation-rich projects, while chain-of-thought prompts perform best on documentation-lean projects. Within the specific scope of LEED v4.1 BD+C compliance verification over raw project documentation, this pipeline and its baselines provide an initial reproducible reference point for both accuracy and failure modes.

Aritro De, Juliana Felkner

The University of Texas at Austin · The University of Texas at Austin,, Austin 78712, Texas, United States of America

LLM Evaluation Neuro-Symbolic AI

Jul 16, 2026cs.AI

Can We Trust Item Response Theory for AI Evaluation?

AI benchmarks increasingly leverage item-level statistical models, particularly item response theory (IRT), to estimate model capabilities, rank systems, select informative examples, and diagnose benchmark quality. However, AI benchmark data often departs from the data regime of human testing, for which standard IRT estimation tools were originally developed: benchmarks typically involve fewer evaluated models, far more items, and capability distributions that may be skewed, clustered, or multimodal. We examine how these regime mismatches challenge the reliability of IRT modeling for AI evaluation. Using item parameters and capability distributions derived from six widely used LLM benchmarks, we simulate response matrices under three common IRT models and compare four estimation tools used in recent benchmark studies: marginal maximum likelihood, Markov chain Monte Carlo, variational inference, and a neural pseudo-Siamese estimator. Across 18,000 simulation conditions, we systematically evaluate computational feasibility, scalability, and the reliability of IRT inferences about model rankings, predicted performance, and item characteristics. Results show that classical estimators can become infeasible in large benchmark settings, whereas scalable estimators can produce unreliable item-level and ranking inferences with small or nonnormally distributed model sets. This study identifies when latent trait models reliably support or risk distorting AI benchmarking claims, and what sample sizes and diagnostics are needed for trustworthy use.

Han Jiang, Sunbeom Kwon, Jinwen Luo +2

1Johns Hopkins University · University of Illinois Urbana-Champaign · University of California, Los Angeles

LLM Evaluation Item Response Theory

Jul 16, 2026cs.CL

Routing Ceilings Are Domain-Independent: Structural Prior Injection in Code Security Vulnerability Detection

Large language models (LLMs) exhibit a well-documented gap between latent capability and consistent activation: the router hypothesis posits that models possess the knowledge to solve a task but lack reliable internal routing to activate it. Prior work in formal mathematical reasoning (SAIR, Cázares 2026) reports that structural priors (cheatsheets) raise in-distribution performance dramatically, yet collapse below the zero-shot baseline out-of-distribution (OOD) -- and that iterative recalibration amplifies rather than corrects the collapse. We test whether this phenomenon is cross-domain by reproducing the SAIR design in source-code security vulnerability detection, evaluating three LLMs (GPT-OSS-120B, Llama-3.3-70B, Gemma-4-31B) across three vulnerability categories (CWE-798, CWE-284, and the non-CWE N+1 anti-pattern) spanning syntactic, contextual, and semantic complexity, then transferring cheatsheet-augmented prompts to real-world CVE data from VUDENC (CWE-89, CWE-22). Our findings replicate and extend SAIR: (F1) structural priors lift semantic-vulnerability recall from 20.0% to 100.0% across all models; (F2) zero-shot performance degrades along a semantic complexity gradient; (F3) the same cheatsheets that saturate synthetic performance amplify distribution-shift collapse on real CVE data (CWE-89: 100% synthetic F1 to 48.9% on VUDENC, -51.1pp); (F5) iterative recalibration produces a v2 cheatsheet that performs worse than v1 on real data, mirroring SAIR's AN45c-vs-AN38 finding. These results provide evidence that the cross-distribution trade-off surface documented in SAIR generalises to code security, and that the router hypothesis is cross-domain. We argue the structural nature of the collapse motivates distribution-aware training over prompt calibration. Code and evaluation scripts: https://github.com/bytepro-ai/bitcoder-v2-research

Manuel Israel Cázares

Bytepro AI Mazatlán, Sinaloa, Mexico

LLM Evaluation Software Vulnerability Detection OOD Generalization In-Context Learning Language Model Robustness

Jul 16, 2026cs.CL

How Well Does AI-Generated Feedback Work? Intrinsic and Extrinsic Evaluation across more than 20,000 EFL Essay Drafts

This study examines feedback in English as a Foreign Language (EFL) writing contexts, focusing on written corrective feedback (WCF). Large language models (LLMs) can provide WCF at scale, but aligning them with pedagogical best practices remains an ongoing challenge. WCF meeting criteria like factuality or relevance may still be unsuitable for learning contexts, highlighting the need for extrinsic evaluation based on the learner's perspective. We deployed WCF systems in a university-level EFL class with nearly 2,000 students, collecting over 20,000 drafts. We evaluated the generated WCF from two perspectives: intrinsic evaluation by experienced English teachers using a rubric, and extrinsic evaluation via student feedback and engagement metrics. Results revealed low alignment between teacher expert ratings and student feedback. These findings suggest that traditional expert evaluation alone may not fully capture WCF's usability or helpfulness from the learner's perspective, highlighting the importance of learner-centered evaluation frameworks for AI-based applications in language education.

Steven Coyne, Diana Galvan-Sosa, Ryan Spring +4

Tohoku University, Sendai, Japan · RIKEN, Wako, Japan · ALTA Institute, Computer Laboratory, University of Cambridge, Cambridge, UK +2

LLM Evaluation Educational Assessment

Jul 16, 2026cs.CL

Controlled Reformulation Testing for Logical Consistency in Large Language Models

Large language models (LLMs) frequently contradict themselves when the surface form of a logically equivalent question changes. We present a benchmark of 350 question families (1,750 total questions) for Controlled Reformulation Testing (CRTBench) to evaluate logical invariance. In this benchmark, we investigate LLMs' ability to maintain consistent answers across controlled reformulations, which include contrapositive rewriting, double negation, negation flipping, and passive voice. We evaluate several frontier LLMs and observe an accuracy-consistency gap where GPT-5.4-mini achieves

98.9\%

base accuracy but only

60.3\%

family-level consistency, while reasoning-optimized o4-mini achieves

96.9\%

consistency. From our experiments, we observe that failures cluster around logically nontrivial transformations such as contrapositive rewriting (

72.4\%

for GPT-5.4-mini) and double negation (

84.6\%

), while surface-level rephrasing remains robust (

94-100\%

). Increasing reasoning effort improves GPT-5.4-mini to

85.4\%

consistency, but leaves GPT-5.4 unchanged overall because gains on nested negation are offset by failures on quantifier families. These results show that accuracy alone is not enough for evaluating logical reasoning in LLMs.

Alexander Gu, Alan Chen

University of Texas at Austin

LLM Evaluation Reasoning Consistency in Language Models Language Model Robustness Self-Consistency in Language Models LLM Reasoning

Jul 15, 2026cs.AI

Instrument Effects in Language-Model Honesty Evaluation: An Auditable Single-System Demonstration

Evaluations of language-model honesty read the model's verdicts as evidence about the model. We test the instrument instead. We built a text-adventure world where the game engine, not any model, knows whether the quest can be completed. A language model plays under a budget and must eventually declare its quest complete, unreachable, or not yet decidable; the engine scores every verdict. Decision rules were recorded before results were read, and run artifacts bind the revisions they executed; the strength of preregistration varies by series and is disclosed. With the player held fixed, instrument choices substantially changed measured behavior. On four byte-identical anchors, expanding a two-verdict grammar to three verdicts moved strong claims from 38/40 to 7/40, while the new incomplete verdict took 28/40 outcomes; across series 2, 93/158 valid games ended incomplete. One sentence disclosing the success criterion took matched-instance false verdicts from 18/59 to 0/58, through fewer decision points and cleaner decisions. Repeated runs of one fixed configuration produced non-stable verdict distributions on 3 of 4 instances: single runs report samples as dispositions. A formally preregistered narrative-register gradient was falsified; two post-hoc, hypothesis-generating patterns remain: register presence roughly doubled strong claims, and budget rendering moved verdicts more than register content (.383 meter vs .150 lantern). The narrator compressed abundant budgets toward scarcity landmarks, yet the registered mediation test returned a null. We propose a four-check integrity protocol for eval instruments.

Justin Bronder

Corabo Inc.

LLM Evaluation LLM Honesty