LLM Evaluation

LLM: Large Language Model

Ask about this topic

Overview

The field studies how to measure language-model capabilities, reliability, and risks, and whether those measurements support sound comparisons and deployment decisions. Approaches include curated and synthetic benchmarks, psychometric and adaptive testing, execution-based checks, robustness tests, and human or LLM-based judges. Researchers also examine calibration, uncertainty, contamination, prompt sensitivity, and whether scores reflect the intended capability across languages and domains. A central tension is that efficient aggregate scores can hide important errors and measurement biases, while more realistic, human-grounded evaluations are costly and difficult to standardize.

Latest papers

CardsList

2,432 papers286 papers in the last four weeks, up 16% on the four weeks before. 1.7% of all new papers.

Mar 24, 2026

Benchmarking Multi-Agent LLM Architectures for Financial Document Processing: A Comparative Study of Orchestration Patterns, Cost-Accuracy Tradeoffs and Production Scaling Strategies

The adoption of large language models (LLMs) for structured information extraction from financial documents has accelerated rapidly, yet production deployments face fundamental architectural decisions with limited empirical guidance. We present a systematic benchmark comparing four multi-agent orchestration architectures: sequential pipeline, parallel fan-out with merge, hierarchical supervisor-worker and reflexive self-correcting loop. These are evaluated across five frontier and open-weight LLMs on a corpus of 10,000 SEC filings (10-K, 10-Q and 8-K forms). Our evaluation spans 25 extraction field types covering governance structures, executive compensation and financial metrics, measured along five axes: field-level F1, document-level accuracy, end-to-end latency, cost per document and token efficiency. We find that reflexive architectures achieve the highest field-level F1 (0.943) but at 2.3x the cost of sequential baselines, while hierarchical architectures occupy the most favorable position on the cost-accuracy Pareto frontier (F1 0.921 at 1.4x cost). We further present ablation studies on semantic caching, model routing and adaptive retry strategies, demonstrating that hybrid configurations can recover 89% of the reflexive architecture's accuracy gains at only 1.15x baseline cost. Our scaling analysis from 1K to 100K documents per day reveals non-obvious throughput-accuracy degradation curves that inform capacity planning. These findings provide actionable guidance for practitioners deploying multi-agent LLM systems in regulated financial environments.

Siddhant Kulkarni, Yukta Kulkarni

Department of Computer Science and Engineering, New York University, New York, NY 10012

Multi-Agent LLM SystemsLLM Agent OrchestrationDocument Information ExtractionLLM EvaluationFinancial Services
AskPDF
Mar 23, 2026

Understanding LLM Performance Degradation in Multi-Instance Processing: The Roles of Instance Count and Context Length

Users often rely on Large Language Models (LLMs) for processing multiple documents or performing analysis over a number of instances. For example, analysing the overall sentiment of a number of movie reviews requires an LLM to process the sentiment of each review individually in order to provide a final aggregated answer. While LLM performance on such individual tasks is generally high, there has been little research on how LLMs perform when dealing with multi-instance inputs. In this paper, we perform a comprehensive evaluation of the multi-instance processing (MIP) ability of LLMs for tasks in which they excel individually. The results show that all LLMs follow a pattern of slight performance degradation for small numbers of instances (approximately 20-100), followed by a performance collapse on larger instance counts. Crucially, our analysis shows that while context length is associated with this degradation, the number of instances has a stronger effect on the final results. This finding suggests that when optimising LLM performance for MIP, attention should be paid to both context length and, in particular, instance count.

Jingxuan Chen, Mohammad Taher Pilehvar, Jose Camacho-Collados

School of Computer Science and Informatics, Cardiff University

LLM EvaluationLong-Context Language Model Evaluation
AskPDF
Mar 23, 2026

Can Large Language Models Detect Methodological Flaws? Evidence from Gesture Recognition for UAV-Based Rescue Operation Based on Deep Learning

Reliable evaluation is essential in machine learning research, yet methodological flaws-particularly data leakage-continue to undermine the validity of reported results. In this work, we investigate whether large language models (LLMs) can act as independent analytical agents capable of identifying such issues in published studies. As a case study, we analyze a gesture-recognition paper reporting near-perfect accuracy on a small, human-centered dataset. We first show that the evaluation protocol is consistent with subject-level data leakage due to non-independent training and test splits. We then assess whether this flaw can be detected independently by six state-of-the-art LLMs, each analyzing the original paper without prior context using an identical prompt. All models consistently identify the evaluation as flawed and attribute the reported performance to non-independent data partitioning, supported by indicators such as overlapping learning curves, minimal generalization gap, and near-perfect classification results. These findings suggest that LLMs can detect common methodological issues based solely on published artifacts. While not definitive, their consistent agreement highlights their potential as complementary tools for improving reproducibility and supporting scientific auditing.

Domonkos Varga

LLM AuditingLLM EvaluationData Leakage
AskPDF
Mar 23, 2026

Silicon Bureaucracy and AI Test-Oriented Education: Contamination Sensitivity and Score Confidence in LLM Benchmarks

Public benchmarks increasingly govern how large language models (LLMs) are ranked, selected, and deployed. We frame this benchmark-centered regime as Silicon Bureaucracy and AI Test-Oriented Education, and argue that it rests on a fragile assumption: that benchmark scores directly reflect genuine generalization. In practice, however, such scores may conflate exam-oriented competence with principled capability, especially when contamination and semantic leakage are difficult to exclude from modern training pipelines. We therefore propose an audit framework for analyzing contamination sensitivity and score confidence in LLM benchmarks. Using a router-worker setup, we compare a clean-control condition with noisy conditions in which benchmark problems are systematically deleted, rewritten, and perturbed before being passed downstream. For a genuinely clean benchmark, noisy conditions should not consistently outperform the clean-control baseline. Yet across multiple models, we find widespread but heterogeneous above-baseline gains under noisy conditions, indicating that benchmark-related cues may be reassembled and can reactivate contamination-related memory. These results suggest that similar benchmark scores may carry substantially different levels of confidence. Rather than rejecting benchmarks altogether, we argue that benchmark-based evaluation should be supplemented with explicit audits of contamination sensitivity and score confidence.

Yiliang Song, Hongjun An, Jiangan Chen +4

Institute of Artificial Intelligence (TeleAI), China Telecom · Guangxi Normal University · Northwestern Polytechnical University

LLM EvaluationBenchmark ContaminationBenchmark ValidityData Contamination in Language ModelsLLM Auditing
AskPDF
Mar 22, 2026

PROMPT2BOX:Improving LLM Weakness Discovery and Specificity Estimation by Uncovering Entailment Structure among Prompts

To discover the weaknesses of LLMs, researchers often embed prompts into a vector space and cluster them to extract insightful patterns. However, vector embeddings primarily capture topical similarity; as a result, prompts that share a topic but differ in specificity, and consequently in difficulty, are often represented similarly, making fine-grained weakness analysis difficult. To address this limitation, we propose Prompt2Box, which embeds prompts into a box embedding space using a trained encoder. The encoder, trained on existing and synthesized datasets, outputs box embeddings that capture not only semantic similarity but also specificity relations between prompts (e.g., "writing an adventure story" is more specific than "writing a story"). We further develop a novel dimension reduction technique for box embeddings to facilitate dataset visualization and comparison. Our experiments demonstrate that box embeddings consistently capture prompt specificity better than vector baselines and achieve 45% error reduction on average in predicting specificity compared to the prompt length baseline. On the downstream task of creating hierarchical clustering trees for 17 LLMs from the UltraFeedback dataset, Prompt2Box can identify 13.5% more LLM weaknesses than vector baselines and achieves an approximately 33% stronger correlation between hierarchical depth and instruction specificity. The code is available at https://github.com/zawedcvg/box_embeddings.

Neeladri Bhuiya, Shib Sankar Dasgupta, Andrew McCallum +1

CICS, University of Massachusetts Amherst, USA · A10 Networks · Amazon AWS AI

LLM EvaluationRepresentation LearningLanguage Model Probing
AskPDF
Mar 22, 2026

Task-Specific Efficiency Analysis: When Small Language Models Outperform Large Language Models

Large Language Models achieve remarkable performance but incur substantial computational costs unsuitable for resource-constrained deployments. This paper presents the first comprehensive task-specific efficiency analysis comparing 16 language models across five diverse NLP tasks. We introduce the Performance-Efficiency Ratio (PER), a novel metric integrating accuracy, throughput, memory, and latency through geometric mean normalization. Our systematic evaluation reveals that small models (0.5--3B parameters) achieve superior PER scores across all given tasks. These findings establish quantitative foundations for deploying small models in production environments prioritizing inference efficiency over marginal accuracy gains.

Jinghan Cao, Yu Ma, Xinjin Li +2

San Francisco State University - Department of Computer Science 1600 Holloway Ave, San Francisco - USA · Carnegie Mellon University - Department of Computer Science 5000 Forbes Ave, Pittsburgh - USA · Columbia University - Department of Computer Science 116th and Broadway, New York - USA +2

Small Language ModelsEfficient Language Model InferenceLLM Evaluation
AskPDF
Mar 22, 2026

COINBench: Moving Beyond Individual Perspectives to Collective Intent Understanding

Understanding human intent is a high-level cognitive challenge for Large Language Models (LLMs), requiring sophisticated reasoning over noisy, conflicting, and non-linear discourse. While LLMs excel at following individual instructions, their ability to distill Collective Intent - the process of extracting consensus, resolving contradictions, and inferring latent trends from multi-source public discussions - remains largely unexplored. To bridge this gap, we introduce COIN-BENCH, a dynamic, real-world, live-updating benchmark specifically designed to evaluate LLMs on collective intent understanding within the consumer domain. Unlike traditional benchmarks that focus on transactional outcomes, COIN-BENCH operationalizes intent as a hierarchical cognitive structure, ranging from explicit scenarios to deep causal reasoning. We implement a robust evaluation pipeline that combines a rule-based method with an LLM-as-the-Judge approach. This framework incorporates COIN-TREE for hierarchical cognitive structuring and retrieval-augmented verification (COIN-RAG) to ensure expert-level precision in analyzing raw, collective human discussions. An extensive evaluation of 20 state-of-the-art LLMs across four dimensions - depth, breadth, informativeness, and correctness - reveals that while current models can handle surface-level aggregation, they still struggle with the analytical depth required for complex intent synthesis. COIN-BENCH establishes a new standard for advancing LLMs from passive instruction followers to expert-level analytical agents capable of deciphering the collective voice of the real world. See our project page on COIN-BENCH.

Xiaozhe Li, Tianyi Lyu, Siyi Yang +6

Tongji University · CurrentsAI Research · Stanford University

LLM Evaluation
AskPDF
Mar 21, 2026

Errors in AI-Assisted Retrieval of Medical Literature: A Comparative Study

Large language models (LLMs) assisted literature retrieval may lead to erroneous references, but these errors have not been rigorously quantified. Therefore, we quantitatively assess errors in reference retrieval of widely used free-version LLM platforms and identify the factors associated with retrieval errors. We evaluated 2,000 references retrieved by 5 LLMs (Grok-2, ChatGPT GPT-4.1, Google Gemini Flash 2.5, Perplexity AI, and DeepSeek GPT-4) for 40 randomly-selected original articles (10 per journal) published Jan. 2024 to July 2025 from British Medical Journal (BMJ), Journal of the American Medical Association, and The New England Journal of Medicine (NEJM). Primary outcomes were a multimetric score ratio combining validity of digital object identifier, PubMed ID, Google-Scholar link, and relevance; and complete miss rate (proportion of references failing all applicable metrics). Multivariable regression was used to examine independent associations. LLM platforms completely failed to retrieve correct reference data 47.8% of the time. The average score ratio of the 5 LLM platforms was 0.29 (standard deviation, 0.35; range, 0-1.25), with a higher score ratio indicating a higher accuracy in retrieving relevant references and correct bibliographic data. The highest and lowest accuracies were achieved by Grok (0.57) and Genimi (0.11), respectively. Compared with BMJ, NEJM articles had lower score ratios and higher complete miss rates. Multivariable analysis shows LLM platforms and journals were independently associated with score ratios and complete miss rate, respectively. We show modest overall performance of LLMs and significant variability in retrieval accuracy across platforms and journals. LLM platforms and journals are associated with LLM's performance in retrieving medical literature. Bibliographic data should be carefully reviewed when using LLM-assisted literature retrieval.

Jenny Gao, Yongfeng Zhang, Mary L Disis +1

College of Arts and Science, New York University, New York, NY · College of Arts and Science, New York University, New York, NY 10003 · Department of Computer Sciences, School of Arts & Sciences, Rutgers University, Piscataway, NJ +9

LLM ReliabilityBiomedical IRInformation RetrievalLLM EvaluationIR Evaluation
AskPDF
Mar 21, 2026

RubricRAG: Towards Interpretable and Reliable LLM Evaluation via Domain Knowledge Retrieval for Rubric Generation

Large language models (LLMs) are increasingly evaluated and sometimes trained using automated graders such as LLM-as-judges that output scalar scores or preferences. While convenient, these approaches are often opaque: a single score rarely explains why an answer is good or bad, which requirements were missed, or how a system should be improved. This lack of interpretability limits their usefulness for model development, dataset curation, and high-stakes deployment. Query-specific rubric-based evaluation offers a more transparent alternative by decomposing quality into explicit, checkable criteria. However, manually designing high-quality, query-specific rubrics is labor-intensive and cognitively demanding and not feasible for deployment. While previous approaches have focused on generating intermediate rubrics for automated downstream evaluation, it is unclear if these rubrics are both interpretable and effective for human users. In this work, we investigate whether LLMs can generate useful, instance-specific rubrics as compared to human-authored rubrics, while also improving effectiveness for identifying good responses. Through our systematic study on two rubric benchmarks, and on multiple few-shot and post-training strategies, we find that off-the-shelf LLMs produce rubrics that are poorly aligned with human-authored ones. We introduce a simple strategy, RubricRAG, which retrieves domain knowledge via rubrics at inference time from related queries. We demonstrate that RubricRAG can generate more interpretable rubrics both for similarity to human-authored rubrics, and for improved downstream evaluation effectiveness. Our results highlight both the challenges and a promising approach of scalable, interpretable evaluation through automated rubric generation.

Kaustubh D. Dhole, Eugene Agichtein

Department of Computer Science, Emory University, Atlanta, GA, USA

Automatic Rubric GenerationLLM EvaluationRubric-Based EvaluationRubric-Guided GenerationRAG Evaluation
AskPDF
Mar 21, 2026

Can ChatGPT Really Understand Modern Chinese Poetry?

ChatGPT has demonstrated remarkable capabilities on both poetry generation and translation, yet its ability to truly understand poetry remains unexplored. Previous poetry-related work merely analyzed experimental outcomes without addressing fundamental issues of comprehension. This paper introduces a comprehensive framework for evaluating ChatGPT's understanding of modern poetry. We collaborated with professional poets to evaluate ChatGPT's interpretation of modern Chinese poems by different poets along multiple dimensions. Evaluation results show that ChatGPT's interpretations align with the original poets' intents in over 73% of the cases. However, its understanding in certain dimensions, particularly in capturing poeticity, proved to be less satisfactory. These findings highlight the effectiveness and necessity of our proposed framework. This study not only evaluates ChatGPT's ability to understand modern poetry but also establishes a solid foundation for future research on LLMs and their application to poetry-related tasks.

Shanshan Wang, Derek F. Wong, Jingming Yao +1

NLP2CT Lab, Department of Computer and Information Science, University of Macau · Department of Portuguese, Faculty of Arts and Humanities, University of Macau

LLM Evaluation
AskPDF
Mar 21, 2026

BenchBench: Benchmarking Automated Benchmark Generation

Benchmarks are the de facto standard for tracking progress in large language models (LLMs), yet static test sets can rapidly saturate, become vulnerable to contamination, and are costly to refresh. Scalable evaluation of open-ended items often relies on LLM judges, introducing additional sources of bias and prompt sensitivity. We argue that evaluation must extend beyond how well models answer benchmarks to how well models design them. We introduce BenchBench, a three-stage pipeline and dataset for benchmarking automated benchmark generation: (i) extract structured domain cards from seed benchmarks, (ii) prompt multiple designer LLMs to generate quota-controlled suites, and (iii) validate items with a multi-model answerer panel using exact/numeric/symbolic verifiers when possible and rubric-guided judging otherwise, yielding designer--answerer matrices with item-level quality flags and psychometric diagnostics. Across nine variants spanning computer science, mathematics, medicine, and theory-of-mind reasoning (including multilingual and multimodal settings), we generate 16.7K items, retain ~15K core items post-filtering, and produce ~152K graded model--item responses. BenchBench shows that benchmark-design ability is only moderately correlated with answer-time strength (Spearman rho 0.37), invalidity is negatively associated with discrimination (Pearson r0.62), and the resulting designer--answerer matrices enable scalable audits of format/modality/language fidelity and suite-dependent self/family interactions. The project is available at: https://github.com/koanatakiyo/BenchBench.

Yandan Zheng, Haoran Luo, Zhenghong Lin +2

Nanyang Technological University Singapore, Singapore

Benchmark DesignSynthetic Benchmark GenerationLLM EvaluationBenchmark Validity
AskPDF
Mar 21, 2026

RLVR Training of LLMs Does Not Improve Thinking Ability for General QA: Evaluation Method and a Simple Solution

Reinforcement learning from verifiable rewards (RLVR) stimulates the thinking processes of large language models (LLMs), substantially enhancing their reasoning abilities on verifiable tasks. It is often assumed that similar gains should transfer to general question answering (GQA), but this assumption has not been thoroughly validated. To assess whether RLVR automatically improves LLM performance on GQA, we propose a Cross-Generation evaluation framework that measures the quality of intermediate reasoning by feeding the generated thinking context into LLMs of varying capabilities. Our evaluation leads to a discouraging finding: the efficacy of the thinking process on GQA tasks is markedly lower than on verifiable tasks, suggesting that explicit training on GQA remains necessary in addition to training on verifiable tasks. We further observe that direct RL training on GQA is less effective than RLVR. Our hypothesis is that, whereas verifiable tasks demand robust logical chains to obtain high rewards, GQA tasks often admit shortcuts to high rewards without cultivating high-quality thinking. To avoid possible shortcuts, we introduce a simple method, Separated Thinking And Response Training (START), which first trains only the thinking process, using rewards defined on the final answer. We show that START improves both the quality of thinking and the final answer across several GQA benchmarks and RL algorithms.

Kaiyuan Li, Jing-Cheng Pang, Yang Yu

National Key Laboratory for Novel Software Technology & School of Artificial Intelligence, Nanjing University, China · Polixir.ai, China · Huawei Technologies Co., Ltd., China

Language Model Reasoning EvaluationRL for Language Model ReasoningReinforcement Learning with Verifiable RewardsQuestion AnsweringLLM Evaluation
AskPDF
Mar 20, 2026

Large Language Models for Missing Data Imputation: Understanding Behavior, Hallucination Effects, and Control Mechanisms

Data imputation is a cornerstone technique for handling missing values in real-world datasets, which are often plagued by missingness. Despite recent progress, prior studies on Large Language Models-based imputation remain limited by scalability challenges, restricted cross-model comparisons, and evaluations conducted on small or domain-specific datasets. Furthermore, heterogeneous experimental protocols and inconsistent treatment of missingness mechanisms (MCAR, MAR, and MNAR) hinder systematic benchmarking across methods. This work investigates the robustness of Large Language Models for missing data imputation in tabular datasets using a zero-shot prompt engineering approach. To this end, we present a comprehensive benchmarking study comparing five widely used LLMs against six state-of-the-art imputation baselines. The experimental design evaluates these methods across 29 datasets (including nine synthetic datasets) under MCAR, MAR, and MNAR mechanisms, with missing rates of up to 20%. The results demonstrate that leading LLMs, particularly Gemini 3.0 Flash and Claude 4.5 Sonnet, consistently achieve superior performance on real-world open-source datasets compared to traditional methods. However, this advantage appears to be closely tied to the models' prior exposure to domain-specific patterns learned during pre-training on internet-scale corpora. In contrast, on synthetic datasets, traditional methods such as MICE outperform LLMs, suggesting that LLM effectiveness is driven by semantic context rather than purely statistical reconstruction. Furthermore, we identify a clear trade-off: while LLMs excel in imputation quality, they incur significantly higher computational time and monetary costs. Overall, this study provides a large-scale comparative analysis, positioning LLMs as promising semantics-driven imputers for complex tabular data.

Arthur Dantas Mangussi, Ricardo Cardoso Pereira, Ana Carolina Lorena +1

Computer Science Division, Aeronautics Institute of Technology, Praça Marechal Eduardo Gomes, 50, São José dos Campos, 12228-900, São Paulo, Brazil · Science and Technology Institute, Federal University of São Paulo, Avenue Cesare Monsueto Giulio Lattes, 1201, São José dos Campos, 12231-280, São Paulo, Brazil · University of Coimbra, CISUC/LASI – Centre for Informatics and Systems of the University of Coimbra, Department of Informatics Engineering, Polo II, Pinhal de Marrocos, Coimbra 3030-290, Portugal

Incomplete Data ImputationLLM Evaluation
AskPDF
Mar 20, 2026

Evaluating Large Language Models on Historical Health Crisis Knowledge in Resource-Limited Settings: A Hybrid Multi-Metric Study

Large Language Models (LLMs) offer significant potential for delivering health information. However, their reliability in low-resource contexts remains uncertain. This study evaluates GPT-4, Gemini Pro, Llama~3, and Mistral-7B on health crisis-related enquiries concerning COVID-19, dengue, the Nipah virus, and Chikungunya in the low-resource context of Bangladesh. We constructed a question--answer dataset from authoritative sources and assessed model outputs through semantic similarity, expert-model cross-evaluation, and Natural Language Inference (NLI). Findings highlight both the strengths and limitations of LLMs in representing epidemiological history and health crisis knowledge, underscoring their promise and risks for informing policy in resource-constrained environments.

Mohammed Rakibul Hasan

Department of Electrical & Computer Engineering, North South University, Dhaka, Bangladesh

LLM ReliabilityLLM EvaluationMedical LLMsFactual Knowledge in Language Models
AskPDF
Mar 20, 2026

Policies Permitting LLM Use for Polishing Peer Reviews Are Currently Not Enforceable

A number of scientific conferences and journals have recently enacted policies that prohibit LLM usage by peer reviewers, except for polishing, paraphrasing, and grammar correction of otherwise human-written reviews. But, are these policies enforceable? To answer this question, we assemble a dataset of peer reviews simulating multiple levels of human-AI collaboration, and evaluate five state-of-the-art detectors, including two commercial systems. Our analysis shows that all detectors misclassify a non-trivial fraction of LLM-polished reviews as AI-generated, thereby risking false accusations of academic misconduct. We further investigate whether peer-review-specific signals, including access to the paper manuscript and the constrained domain of scientific writing, can be leveraged to improve detection. While incorporating such signals yields measurable gains in some settings, we identify limitations in each approach and find that none meets the accuracy standards required for identifying AI use in peer reviews. Importantly, our results suggest that recent public estimates of AI use in peer reviews through the use of AI-text detectors should be interpreted with caution, as current detectors misclassify mixed reviews (collaborative human-AI outputs) as fully AI generated, potentially overstating the extent of policy violations.

Rounak Saha, Gurusha Juneja, Dayita Chaudhuri +3

♠Indian Institute of Science · ♢University of California, Santa Barbara · ♡Carnegie Mellon University

LLM EvaluationAI-Generated Text DetectionScholarly Peer Review
AskPDF
Mar 20, 2026

When Demonstrations Fail: Diagnosing the Limits of In-Context Learning in Large Audio-Language Models with Progressive Cue Removal

While Large Audio-Language Models (LALMs) have been shown to exhibit degraded instruction-following capabilities, their ability to infer task patterns from in-context examples with audio remains understudied. To address this gap, we design a three-stage evaluation pipeline that progressively reduces textual guidance to systematically evaluate LALMs' in-context learning ability in the audio modality. Evaluating six LALMs across four audio understanding tasks under two output constraint categories, we uncover a consistent asymmetry across LALMs: in-context demonstrations reliably improve format compliance but fail to improve the core task performance. This suggests that LALMs can glean surface-level formatting patterns from demonstrations but may struggle to leverage cross-modal semantic grounding to reliably infer task objectives from examples with audio, highlighting potential limitations in current cross-modal integration. We further probe how demonstrations are used through two complementary analyses, demonstration label shuffling and attention knockout on demonstration spans, both showing that LALMs leverage in-context examples primarily to establish the output label space and format rather than to learn a meaningful input-output correspondence.

Yen-Ting Piao, Jay Chiehen Liao, Wei-Tang Chien +5

National Taiwan University, Taiwan

Audio-Language Model EvaluationLLM EvaluationLLM GroundingMultimodal ICLAudio UnderstandingAudio-Language Models+1
AskPDF
Mar 20, 2026

Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluation

Recent work on chain-of-thought (CoT) faithfulness reports single aggregate numbers (e.g., DeepSeek-R1 acknowledges hints 39% of the time), implying that faithfulness is an objective, measurable property of a model. This paper provides evidence that it is not. Three classifiers (a regex-only detector, a regex-plus-LLM pipeline, and a Claude Sonnet 4 judge) are applied to 10,276 influenced reasoning traces from 12 open-weight models spanning 9 families and 7B to 1T parameters. On identical data, these classifiers produce faithfulness rates of 74.4%, 82.6%, and 69.7%. Per-model gaps range from 2.6 to 30.6 percentage points; all pairwise McNemar tests are significant (p < 0.001). The disagreements are systematic: Cohen's kappa ranges from 0.06 ("slight") for sycophancy hints to 0.42 ("moderate") for grader hints, and the asymmetry is pronounced: for sycophancy, 883 cases are classified as faithful by the pipeline but unfaithful by the Sonnet judge, while only 2 go the other direction. Classifier choice can also reverse model rankings: Qwen3.5-27B ranks 1st under the pipeline but 7th under Sonnet; OLMo-3.1-32B moves from 9th to 3rd. Different classifiers operationalize faithfulness at different levels of stringency (lexical mention versus epistemic dependence), yielding divergent measurements on the same behavior. These results indicate that published faithfulness numbers cannot be meaningfully compared across studies using different classifiers, and that future evaluations should report sensitivity ranges across multiple classification methodologies.

Richard J. Young

University of Nevada, Las Vegas, Department of Management, Entrepreneurship and Technology, Lee Business School, Las Vegas, NV, USA · DeepNeuro AI, Las Vegas, NV, USA

CoT FaithfulnessLLM EvaluationFaithfulness of Language Model Explanations
AskPDF
Mar 20, 2026

Evaluating AI-based Scientific Knowledge Synthesis with Epidemiological Systematic Reviews

Systematic literature reviews (SLRs) are a demanding and high-stakes form of scientific knowledge synthesis that remains underspecified as an evaluation setting for large language models (LLMs). We introduce AgentSLR, a large-scale evaluation harness comprising an SLR automation workflow and an expert annotated dataset covering 16,248 articles, designed to test LLM capabilities across the stages of SLRs in epidemiology. Reference annotations were derived from peer-reviewed studies on WHO priority pathogens and produced by domain experts. The harness evaluates each review stage as a separate unit with dedicated metrics enabling targeted failure analysis. We evaluated five frontier reasoning models and found that no single model dominated across all tasks, showing sub-task specialisation often hidden by aggregate benchmarks. Structured data extraction is a major bottleneck, with no model exceeding an average field-level F1 of 0.67. Estimated costs vary substantially, by up to 96 times across evaluated models. Documented failure modes suggest that the evaluated models are not yet reliable enough for unsupervised deployment in epidemiology, where findings can inform public policy.

Shreyansh Padarha, Ryan Othniel Kearns, Tristan Naidoo +13

University of Oxford · Imperial College London · University of Nottingham +2

Systematic Review AutomationLLM EvaluationScientific Workflow AutomationLLM Information Extraction
AskPDF
Mar 20, 2026

EVE: A Domain-Specific LLM Framework for Earth Intelligence

We introduce Earth Virtual Expert (EVE), the first open-source, end-to-end initiative for developing and deploying domain-specialized LLMs for Earth Intelligence. At its core is EVE-Instruct, a domain-adapted 24B model built on Mistral Small 3.2 and optimized for reasoning and question answering. On newly constructed Earth Observation and Earth Sciences benchmarks, it outperforms comparable models while preserving general capabilities. We release curated training corpora and the first systematic domain-specific evaluation benchmarks, covering MCQA, open-ended QA, and factuality. EVE further integrates RAG and a hallucination-detection pipeline into a production system deployed via API and GUI, supporting 350 pilot users so far. All models, datasets, and code are ready to be released under open licenses as contributions to our field at huggingface.co/eve-esa and github.com/eve-esa.

Àlex R. Atrio, Antonio Lopez, Jino Rohit +6

Pi School · Mistral AI · Translated +1

Earth ScienceLLM Evaluation
AskPDF
Mar 20, 2026

SEAR: Schema-Based Evaluation and Routing for LLM Gateways

Evaluating production LLM responses and routing requests across providers in LLM gateways requires fine-grained quality signals and operationally grounded decisions. To address this gap, we present SEAR, a schema-based evaluation and routing system for multi-model, multi-provider LLM gateways. SEAR defines an extensible relational schema covering both LLM evaluation signals (context, intent, response characteristics, issue attribution, and quality scores) and gateway operational metrics (latency, cost, throughput), with cross-table consistency links across around one hundred typed, SQL-queryable columns. To populate the evaluation signals reliably, SEAR proposes self-contained signal instructions, in-schema reasoning, and multi-stage generation that produces database-ready structured outputs. Because signals are derived through LLM reasoning rather than shallow classifiers, SEAR captures complex request semantics, enables human-interpretable routing explanations, and unifies evaluation and routing in a single query layer. Across thousands of production sessions, SEAR achieves strong signal accuracy on human-labeled data and supports practical routing decisions, including large cost reductions with comparable quality.

Zecheng Zhang, Han Zheng, Yue Xu

Strukto.AI, USA · Infron.AI, USA

LLM RoutingLLM EvaluationLLM ServingStructured Output Generation
AskPDF
Mar 20, 2026

TextReasoningBench: Does Reasoning Really Improve Text Classification in Large Language Models?

Eliciting explicit, step-by-step reasoning traces from large language models (LLMs) has emerged as a dominant paradigm for enhancing model capabilities. Although such reasoning strategies were originally designed for problems requiring explicit multi-step reasoning, they have increasingly been applied to a broad range of NLP tasks. This expansion implicitly assumes that deliberative reasoning uniformly benefits heterogeneous tasks. However, whether such reasoning mechanisms truly benefit classification tasks remains largely underexplored, especially considering their substantial token and time costs. To fill this gap, we introduce TextReasoningBench, a systematic benchmark designed to evaluate the effectiveness and efficiency of reasoning strategies for text classification with LLMs. We compare seven reasoning strategies, namely IO, CoT, SC-CoT, ToT, GoT, BoC, and long-CoT across ten LLMs on five text classification datasets. Beyond traditional metrics such as accuracy and macro-F1, we introduce two cost-aware evaluation metrics that quantify the performance gain per reasoning token and the efficiency of performance improvement relative to token cost growth. Experimental results reveal three notable findings: (1) Reasoning does not universally improve classification performance: while moderate strategies such as CoT and SC-CoT yield consistent but limited gains (typically +1% to +3% on big models), more complex methods (e.g., ToT and GoT) often fail to outperform simpler baselines and can even degrade performance, especially on small models; (2) Reasoning is often inefficient: many reasoning strategies increase token consumption by 10×\times to 100×\times (e.g., SC-CoT and ToT) while providing only marginal performance improvements.

Xinyu Guo, Yazhou Zhang, Jing Qin

School of Computer Science and Technology, Tianjin University, China · School of Nursing, The Hong Kong Polytechnic University, HongKong

Large Language Model-Based Text ClassificationLLM EvaluationEfficient Language Model ReasoningText Classification
AskPDF
Mar 19, 2026

ItinBench: Benchmarking Planning Across Multiple Cognitive Dimensions with Large Language Models

Large language models (LLMs) with advanced cognitive capabilities are emerging as agents for various reasoning and planning tasks. Traditional evaluations often focus on specific reasoning or planning questions within controlled environments. Recent studies have explored travel planning as a medium to integrate various verbal reasoning tasks into real-world contexts. However, reasoning tasks extend beyond verbal reasoning alone, and a comprehensive evaluation of LLMs requires a testbed that incorporates tasks from multiple cognitive domains. To address this gap, we introduce ItinBench, a benchmark that features one task of spatial reasoning, i.e., route optimization, into trip itinerary planning while keeping the traditional verbal reasoning tasks. ItinBench evaluates various LLMs across diverse tasks simultaneously, including Llama 3.1 8B, Mistral Large, Gemini 1.5 Pro, and GPT family. Our findings reveal that LLMs struggle to maintain high and consistent performance when concurrently handling multiple cognitive dimensions. By incorporating tasks from distinct human-level cognitive domains, ItinBench provides new insights into building more comprehensive reasoning testbeds that better reflect real-world challenges. The code and dataset: https://ethanwtl.github.io/IBweb/

Tianlong Wang, Pinqiao Wang, Weili Shi +1

University of Virginia

LLM EvaluationLLM PlanningSpatial Reasoning Benchmarks
AskPDF
Mar 19, 2026

Correct Chains, Wrong Answers: Dissociating Reasoning from Output in LLM Logic

LLMs can execute every step of chain-of-thought reasoning correctly and still produce wrong final answers. We introduce the Novel Operator Test, a benchmark that separates operator logic from operator name, enabling rigorous distinction between genuine reasoning and pattern retrieval. By evaluating Boolean operators under unfamiliar names across depths 1-10 on five models (up to 8,100 problems each), we demonstrate a reasoning-output dissociation that existing benchmarks cannot detect. At Claude Sonnet 4's depth 7, all 31 errors have verifiably correct reasoning yet wrong declared answers; 17/19 errors in mixed-operator chains exhibit the same pattern. The benchmark reveals two failure types: strategy failures at depth 2, where models attempt terse retrieval (+62pp from scaffolding), and content failures at depth 7, where models reason fully but err systematically (+8-30pp, 0/300 errors post-intervention). A Trojan operator (XOR's truth table under a novel name) confirms name alone does not gate reasoning (p >= 0.49), while Llama's novelty gap widens to 28pp at depth 8-9 with the Trojan at 92-100%, isolating genuine difficulty with novel logic from name unfamiliarity.

Abinav Rao, Sujan Rachuri, Nikhil Vemuri

LLM ReasoningLLM EvaluationReasoning Consistency in Language ModelsStructured CoT Reasoning
AskPDF
Mar 19, 2026

FinTradeBench: A Financial Reasoning Benchmark for LLMs

Financial analysis requires reasoning over heterogeneous signals, including company fundamentals from regulatory filings and trading signals derived from market prices. However, existing financial benchmarks focus almost exclusively on regulatory filings, neglecting market dynamics and cross-signal interactions. We introduce FinTradeBench, a benchmark of 1,400 questions covering NASDAQ-100 stocks over a ten-year window. To holistically evaluate financial reasoning, FinTradeBench includes questions in three reasoning categories: fundamentals-focused, trading-signal-focused, and hybrid questions that require cross-signal reasoning. Our benchmark starts with 150 expert-authored seed questions, which are then scaled to 1400. To ensure reliable evaluation, we create a Core400 subset in which every item is expert-reviewed. In zero-shot (No-RAG) response generation, we observe that all models perform substantially worse on trading-focused questions compared to fundamental-focused questions. To identify failure causes, we evaluate LLMs under controlled settings including retrieval-based baselines, tool-assisted computation, oracle conditions, and time-series-specialized pipelines. LLMs' capabilities to do computation and evidence selection contribute most to reasoning failures. Additionally, retrieval significantly improves trading-signal reasoning but does not improve fundamental reasoning, where identifying relevant signals remains the dominant failure mode. Finally, rather than relying on a small, fine-tuned model for time-series reasoning, it is better to summarize the time-series data and leverage LLMs' reasoning capabilities.

Yogesh Agrawal, Aniruddha Dutta, Md Mahadi Hasan +2

University of Central Florida

Financial Reasoning in Language ModelsFinancial Reasoning BenchmarksFinancial QALLM EvaluationNumerical Reasoning in Language ModelsQuantitative Finance
AskPDF
Mar 19, 2026

How Auditory Knowledge in LLM Backbones Shapes Audio Language Models: A Holistic Evaluation

Large language models (LLMs) have been widely used as knowledge backbones of Large Audio Language Models (LALMs), yet how much auditory knowledge they encode through text-only pre-training and how this affects downstream performance remains unclear. We study this gap by comparing different LLMs under two text-only and one audio-grounded setting: (1) direct probing on AKB-2000, a curated benchmark testing the breadth and depth of auditory knowledge; (2) cascade evaluation, where LLMs reason over text descriptions from an audio captioner; and (3) audio-grounded evaluation, where each LLM is fine-tuned into a Large Audio Language Model (LALM) with an audio encoder. Our findings reveal that auditory knowledge varies substantially across families, and text-only results are strongly correlated with audio performance. Our work provides empirical grounding for a comprehensive understanding of LLMs in audio research.

Ke-Han Lu, Szu-Wei Fu, Chao-Han Huck Yang +13

National Taiwan University, Taiwan · NVIDIA · Academia Sinica, Taiwan

Audio UnderstandingAudio-Language Model EvaluationLLM Evaluation
AskPDF
Mar 19, 2026

Evaluating Counterfactual Strategic Reasoning in Large Language Models

We evaluate Large Language Models (LLMs) in repeated game-theoretic settings to assess whether strategic performance reflects genuine reasoning or reliance on memorized patterns. We consider two canonical games, Prisoner's Dilemma (PD) and Rock-Paper-Scissors (RPS), upon which we introduce counterfactual variants that alter payoff structures and action labels, breaking familiar symmetries and dominance relations. Our multi-metric evaluation framework compares default and counterfactual instantiations, showcasing LLM limitations in incentive sensitivity, structural generalization and strategic reasoning within counterfactual environments.

Dimitrios Georgousis, Maria Lymperaiou, Angeliki Dimitriou +2

National Technical University of Athens

Strategic Reasoning Risks in Language ModelsLLM EvaluationCounterfactual Reasoning in Language ModelsCounterfactual Evaluation
AskPDF
Mar 19, 2026

Book your room in the Turing Hotel! A symmetric and distributed Turing Test with multiple AIs and humans

In this paper, we report our experience with TuringHotel'', a novel extension of the Turing Test based on interactions within mixed communities of Large Language Models (LLMs) and human participants. The classical one-to-one interaction of the Turing Test is reinterpreted in a group setting, where both human and artificial agents engage in time-bounded discussions and, interestingly, are both judges and respondents. This community is instantiated in the novel platform UNaIVERSE (https://unaiverse.io), creating a World'' which defines the roles and interaction dynamics, facilitated by the platform's built-in programming tools. All communication occurs over an authenticated peer-to-peer network, ensuring that no third parties can access the exchange. The platform also provides a unified interface for humans, accessible via both mobile devices and laptops, that was a key component of the experience in this paper. Results of our experimentation involving 17 human participants and 19 LLMs revealed that current models are still sometimes confused as humans. Interestingly, there are several unexpected mistakes, suggesting that human fingerprints are still identifiable but not fully unambiguous, despite the high-quality language skills of artificial participants. We argue that this is the first experiment conducted in such a distributed setting, and that similar initiatives could be of national interest to support ongoing experiments and competitions aimed at monitoring the evolution of large language models over time.

Christian Di Maio, Tommaso Guidi, Luigi Quarantiello +4

DIISM, University of Siena, Italy · Department of Computer Science, University of Pisa, Italy · LUISS University, Rome, Italy

LLM EvaluationHuman-AI InteractionMulti-Agent LLMsLarge Language Model-Based Social Simulation
AskPDF
Mar 19, 2026

ZEBRAARENA: A Diagnostic Simulation Environment for Studying Reasoning-Action Coupling in Tool-Augmented LLMs

Tool-augmented large language models (LLMs) must tightly couple multi-step reasoning with external actions, yet existing benchmarks often confound this interplay with complex environment dynamics, memorized knowledge or dataset contamination. In this paper, we introduce ZebraArena, a procedurally generated diagnostic environment for studying reasoning-action coupling in tool-augmented LLMs, with controllable difficulty and a knowledge-minimal design, which limits gains from memorization or dataset contamination. Each task in ZebraArena requires a set of critical information which is available only through targeted tool use, yielding an interpretable interface between external information acquisition and deductive reasoning. This design provides deterministic evaluation via unique solutions, and a theoretical optimal query count for measuring efficient tool use. We show that ZebraArena requires a combination of in-depth reasoning and accurate external tool calling, which remains a challenge as frontier reasoning models such as GPT-5 and Gemini 2.5 Pro only achieves 60% accuracy on the hard instances. We also observe a persistent gaps between theoretical optimality and practical tool usage. For example, GPT-5 uses 70-270% more tool calls than the theoretical optimum. We highlight the key findings in our evaluation, and hope ZebraArena stimulates further research on the interplay between internal reasoning and external action.

Wanjia Zhao, Ludwig Schmidt, Yejin Choi +3

Stanford University · Microsoft Research

LLM EvaluationAI Agent BenchmarksLLM Tool UseLLM Reasoning
AskPDF
Mar 19, 2026

Do Post-Training Algorithms Actually Differ? A Controlled Study Across Model Scales Uncovers Scale-Dependent Ranking Inversions

Post-training alignment has produced dozens of competing algorithms -- DPO, SimPO, KTO, GRPO, and others -- yet practitioners lack controlled comparisons to guide algorithm selection. We present OXRL, a unified framework implementing 51 post-training algorithms with identical infrastructure, enabling the first large-scale apples-to-apples evaluation. Our study spans 8 algorithms across 4 model scales (0.5B--7B), 3 evaluation domains, and a 20-variant DPO taxonomy (100 runs at 1.5B, 5 seeds each), totaling ∼\sim240 training runs on H100 GPUs. Three headline findings emerge. (1)Algorithm rankings are unstable across scale: at 1.5B, online RL (SGRPO) tops all methods at 58.0%±\pm0.57 on GSM8K; by 7B, the worst small-scale method (SimPO) becomes the best (85.8%), a complete ranking inversion driven by model scale rather than LoRA regularization (confirmed via 2×\times2 factorial). (2)Loss function modifications yield negligible gains: none of 20 DPO variants significantly outperform vanilla DPO after Bonferroni correction; the sole significant outlier, SimPO, is worse (−-11.5pp, p<10−4p < 10^{-4}). (3)Algorithm leverage is task-specific: the 19.3pp GSM8K spread collapses to 0.54pp on MATH (36×36\times) and 0.47pp on general-domain benchmarks (41×41\times), confirming that algorithm choice matters primarily within the training distribution. These findings yield a hierarchy of leverage for practitioners: model scale (∼{\sim}50pp) ≫\gg training paradigm (∼{\sim}10pp) ≫\gg online vs.\ offline (∼{\sim}9pp) ≫\gg loss function (∼{\sim}1pp). We release all code, configs, and evaluation data as a living community benchmark.

Xiaoyi Li

Language Model Post-TrainingLLM AlignmentLLM Evaluation
AskPDF
Mar 19, 2026

GAIN: A Benchmark for Goal-Aligned Decision-Making of Large Language Models under Imperfect Norms

We introduce GAIN (Goal-Aligned Decision-Making under Imperfect Norms), a benchmark designed to evaluate how large language models (LLMs) balance adherence to norms against business goals. Existing benchmarks typically focus on abstract scenarios rather than real-world business applications. Furthermore, they provide limited insights into the factors influencing LLM decision-making. This restricts their ability to measure models' adaptability to complex, real-world norm-goal conflicts. In GAIN, models receive a goal, a specific situation, a norm, and additional contextual pressures. These pressures, explicitly designed to encourage potential norm deviations, are a unique feature that differentiates GAIN from other benchmarks, enabling a systematic evaluation of the factors influencing decision-making. We define five types of pressures: Goal Alignment, Risk Aversion, Emotional/Ethical Appeal, Social/Authoritative Influence, and Personal Incentive. The benchmark comprises 1,200 scenarios across four domains: hiring, customer support, advertising and finance. Our experiments show that advanced LLMs frequently mirror human decision-making patterns. However, when Personal Incentive pressure is present, they diverge significantly, showing a strong tendency to adhere to norms rather than deviate from them.

Masayuki Kawarada, Kodai Watanabe, Soichiro Murakami

CyberAgent

LLM Decision-MakingLLM EvaluationNormative Conformity in Language Models
AskPDF