Language Model Generation Evaluation

Ask about this topic

Overview

The field studies how to tell whether generated text is correct, useful, faithful to its sources, well written, and appropriate for a task. Approaches range from lexical and embedding-based metrics to human ratings, rubric-based LLM judges, learned evaluators, and task-specific checks such as code execution or schema validation. Benchmarks increasingly test long-form and open-ended outputs, multiple samples, languages, and settings, while evaluation-cost methods aim to select informative examples efficiently. A central challenge is that scores can depend on prompts, judges, and task context: a fluent or highly rated answer may still omit facts, repeat familiar ideas, or fail to reflect human preferences.

Latest papers

CardsList

396 papers56 papers in the last four weeks, up 17% on the four weeks before. 0.3% of all new papers.

May 24, 2026

Faithfulness Metrics Don't Measure Faithfulness: A Meta-Evaluation with Ground Truth

Chains of thought (CoTs) have become central in interpreting and auditing behaviors of large language models. Yet growing evidence suggests that these traces often fail to faithfully represent the computations behind a model's predictions. Several faithfulness metrics have been proposed, but whether they indeed measure faithfulness remains unknown. Answering this requires ground-truth labels, which are hard to obtain since internal computations are not directly observable. Consequently, most works proposing metrics report only absolute scores or comparisons to prior metrics, and the few existing benchmarks rely on proxies like plausibility or importance, properties orthogonal to faithfulness that can mislead about whether a CoT can be trusted. We address this challenge by constructing tasks whose outputs reveal which intermediate computations must have produced them, and developing an automated labeling pipeline that yields ground-truth faithfulness labels at both the step and CoT level. Building on this methodology, we present BonaFide, a benchmark of 3,066 labeled CoTs across 13 tasks and 10 models, and use it to conduct the first systematic evaluation of prominent faithfulness metrics. Our experiments show that most metrics perform near chance, exhibit strong prediction biases and degrade on longer CoTs. The best metric reaches only 0.70 AUROC at the CoT level while another reaches 0.59 at the step level, with neither transferring across settings, while entailing prohibitively high computational cost. Our results expose fundamental gaps in current faithfulness evaluation and call for the development of more reliable and efficient metrics.

Yoav Gur-Arieh, Ana Marasović, Mor Geva

Tel Aviv University · University of Utah

CoT FaithfulnessLLM EvaluationLanguage Model Generation EvaluationFaithfulness of Language Model ExplanationsLLM Reliability
AskPDF
May 24, 2026

Repeated Sequences Reveal Gaps between Large Language Models and Natural Language

Evaluating whether large language models (LLMs) capture the structure of natural language beyond local fluency remains an open challenge. Existing evaluation methods, largely based on task performance or short-context behavior, provide limited insight into the long-range statistical organization of generated text. We propose a complementary evaluation framework based on repeated subsequences. By analyzing their distribution across scales and relating it to higher-order Rényi entropies, we probe how texts reuse previously established structure under finite-length conditions. Experiments on human-written texts and length-matched GPT-generated texts show that, while power-law models can describe restricted ranges of block length, the observed entropy growth is often equally or better characterized by logarithmic--power forms. Across datasets, natural language exhibits stable entropy-growth patterns over accessible ranges, with consistent average behavior despite variability across individual texts. In contrast, GPT-generated texts show systematic and statistically significant shifts in estimated exponents with model size. These results demonstrate that repeated-subsequence entropy provides a quantitative structural diagnostic that reveals systematic differences in long-range organization, distinguishing natural language from state-of-the-art LLM outputs beyond surface-level fluency.

Kumiko Tanaka-Ishii

Waseda University, Japan.

LLM EvaluationLanguage Model Generation EvaluationSequence Modeling
AskPDF
May 23, 2026

Benchmarking the Personalization Capabilities of Large Language Models

Personalization is classically a two-party problem: a sender chooses what to say, and a receiver with independent objectives decides whether to act. A salesperson pitching the same analytics product leads with HIPAA compliance for a hospital and real-time reporting for a retailer, expecting a different argument to work on each. Existing LLM personalization benchmarks measure a narrower, one-party property: whether output matches the preferences of the same user it serves-sender and receiver being the same, as when RLHF aligns an assistant to its own user. The two-party case is harder to study automatically, since it needs ground truth linking specific content to an observed receiver action. Sales outreach provides this: a message written for one prospect, recorded against whether it produced a reply, a call, or a closed deal. We introduce SDR-Arena, a framework for benchmarking two-party generative personalization at scale, and SDR-Bench, a public corpus of 50,000 customer success stories across 22 industries and 3500 enterprises. Given only pre-outcome information, an agent must reconstruct the arguments that won the deal, scored by a weighted nugget-recall metric (WCS). The best model, Claude Sonnet 4.6, reaches 55.8% WCS indicating it recovers only half the winning content-a plateau we observe across model families that costly deep-research pipelines do not close. An ablation shows the cause is retrieval, not reasoning: models improve substantially given the facts a human researcher would gather, but rarely find them through web search alone. Two studies with professional SDRs support the metric: only 48% of generated pitches were rated usable without editing, and WCS yields model rankings consistent with evaluation against expert strategies authored independently of any model output.

Ashutosh Srivastava, Siddharth Yedlapati, Vinay Aggarwal +4

Adobe Media & Data Science Research

Language Model Generation EvaluationPersonalized Language Models
AskPDF
May 22, 2026

NLG Evaluation: Past, Present, Future

Natural Language Generation (NLG) evaluation has changed dramatically since 1990, and will continue to evolve in the future. In 1990, when NLG had close ties to linguistics, there was very little formal experimental evaluation in the modern sense. In 2026, when NLG is closely linked to machine learning, experimental evaluation is expected and indeed fundamental to research. Many evaluation techniques were developed over this period, including most recently LLM-as-Judge. I expect NLG evaluation will continue to evolve in the future. In particular, impact, qualitative, and safety evaluation will become more important as large numbers of people routinely use NLG technology.

Ehud Reiter

Dept of Computing Science University of Aberdeen Aberdeen, UK

LLM EvaluationLanguage Model Generation EvaluationGenerative AI Evaluation
AskPDF
May 22, 2026

How Human-Like Are Large Language Models? A Register-Aware Linguistic Evaluation Framework

While factual correctness and task-performance have been in focus of Large Language Model (LLM) research for a long time, the fundamental question of how human-like generated texts are on a linguistic level has been underexplored. From a corpus-linguistic perspective, language production is inherently context-dependent, with distinct communicative contexts giving rise to differences in frequencies and co-occurrence patterns of linguistic features. A text failing to adhere to these patterns can be content-wise correct, but still be unfavorable to human readers. In this work, we propose a context-aware evaluation framework in which human-likeness is assessed using a two-sample problem between the linguistic feature distribution of a human reference corpus for a given register and a corresponding LLM-generated corpus. We implement this framework using the Maximum Mean Discrepancy (MMD) and the 67 lexico-grammatical features introduced by Biber, which are commonly applied in corpus linguistics. In our experiments, we compare seven instruction-tuned, open-source models across five English-language datasets spanning distinct registers against a human baseline. While across all tested setups, LLMs deviate from the human baseline, which models are closest to human language depends on the register and is not dictated by model size.

Björn Nieth, Marianna Gracheva, Michaela Mahlberg +2

Department Artificial Intelligence in Biomedical Engineering (AIBE), FAU Erlangen-Nürnberg, Germany · Department of Digital Humanities and Social Studies (DHSS), FAU Erlangen-Nürnberg, Germany · University of Birmingham, United Kingdom +2

LLM EvaluationLanguage Model Generation Evaluation
AskPDF
May 21, 2026

AMEL: Accumulated Message Effects on LLM Judgments

Large language models are routinely used as automated evaluators: to review code, moderate content, or score outputs, often with many items passing through one conversation. We ask whether the polarity of prior conversation history biases subsequent judgments, an effect we call the accumulated message effect on LLM judgments (AMEL). Across 84,088 API calls to 12 models from 5 providers (OpenAI, Anthropic, Google, DeepSeek, and four open-source models), we present identical test items in isolation or following histories saturated with predominantly positive or negative evaluations. Models shift toward the conversation's prevailing polarity (d = -0.17, p < 10^-53). The effect concentrates on items where the model is genuinely uncertain at baseline (d = -0.36 for high-entropy items, vs d = -0.15 when the baseline is deterministic). Bias does not grow with context length: 5 prior turns and 50 produce the same shift (Spearman |r| < 0.01; OLS slope p = 0.80). And there is a negativity asymmetry: paired per item, negative histories induce 1.52x more bias than positive (t = 13.03, p < 10^-36, n = 2,733). Scaling helps but does not solve it (Anthropic: Haiku -0.22 to Opus -0.17; OpenAI: Nano -0.34 to GPT-5.2 -0.17). Three follow-ups narrow the mechanism. The token probability distribution shifts continuously, not at a threshold. The negativity asymmetry has both token-level and semantic components, though attributing the balance is exploratory at our sample sizes. Position does not matter: five biased turns anywhere in a 50-turn history produce the same shift. The simplest fix for evaluation pipelines is a fresh context per item; when batching is unavoidable, balancing the history helps.

Sid-Ali Temkit

LLM EvaluationLanguage Model Generation EvaluationLanguage Model Bias EvaluationBias in LLM-as-a-Judge
AskPDF
May 20, 2026

RankJudge: A Multi-Turn LLM-as-a-Judge Synthetic Benchmark Generator

As interactive LLM-based applications are created and refined, model developers need to evaluate the quality of generated text along many possible axes. For simpler systems, human evaluation may be practical, but in complicated systems like conversational chatbots, the amount of generated text can overwhelm human annotation resources. Model developers have begun to rely heavily on auto-evaluation, where LLMs are also used to judge generation quality. However, existing LLM-as-a-judge benchmarks largely focus on simple Q&A tasks that do not match the complexity of multi-turn conversations. We introduce RankJudge, a benchmark generator for evaluating LLM-as-a-judge on multi-turn conversations grounded in reference documents. RankJudge creates pairs of conversations where one conversation has a single flaw injected into one turn. This construction allows paired conversations to be labeled unambiguously as better or worse, and precisely isolates failure categories to individual turns, enabling a strict joint correctness criterion for judging. We implement RankJudge across the domains of machine learning, biomedicine, and finance, evaluate 21 frontier LLM judges, and rank those judges via the Bradley-Terry model. Our formulation also allows ranking each conversation pair with difficulty ratings, which we use to dynamically curate the evaluation slice to reduce label noise, as confirmed via human annotation. We find that judge rankings are stable under partial observability, coarser correctness criteria, and an alternative random-walk rating algorithm.

Zhenwei Tang, Zhaoyan Liu, Rasa Hosseinzadeh +3

Layer 6 AI · University of Toronto

LLM EvaluationPairwise ComparisonLLM-as-a-JudgeLanguage Model Generation EvaluationPairwise Preference EvaluationPairwise Ranking+2
AskPDF
May 20, 2026

APM: Evaluating Style Personalization in LLMs with Arbitrary Preference Mappings

Typical LLM responses tend to follow a default style, even though users often have distinct preferences regarding tone, verbosity, and formality that they do not explicitly state in their prompts. Evaluating whether personalization methods can adapt to these implicit preferences is challenging, since users typically provide prompts rather than reference responses, style preferences are not factually verifiable, and reference-free LLM judges may conflate personalization with general response quality. To address these challenges, we introduce the Arbitrary Preference Mapping (APM) benchmark, which decouples user attributes (e.g. enthusiastic) from response principles (e.g. persuasive) via a hidden, randomized mapping C\mathbf{C} that maps user attributes to preferences about response traits. Because C\mathbf{C} carries no semantic content and is resampled across runs, models cannot exploit stereotypical associations and must infer preferences from conversation history. Using this unbiased evaluation methodology, we adapt retrieval-augmented, prompt-optimization, and routing personalization methods and evaluate them on Llama-3.1-8B and Qwen-3.5-27B. Our results show that routing is the most reliable approach, while RAG only improves with the stronger base LLM, and soft prompt optimization fails to improve significantly over a non-personalized baseline. Our extensive evaluation reveals that in this realistic setting, personalization remains challenging, but our adapted methods show promise.

Philipp Spohn, Leander Girrbach, Zeynep Akata

Technical University of Munich, Helmholtz Munich · Munich Center for Machine Learning (MCML)

LLM EvaluationLLM PersonalizationLanguage Model Generation EvaluationAutomatic Prompt OptimizationPersonalized Language ModelsLLM Routing+1
AskPDF
May 19, 2026

Less Back-and-Forth: A Comparative Study of Structured Prompting

Large language models (LLMs) are widely used for open-ended tasks, but underspecified prompts can lead to low-quality answers and additional interaction. This paper studies whether structured prompt design improves response quality while reducing user effort. We compare three prompt conditions: a raw prompt, a checklist-improved prompt, and a clarifying-question prompt. We evaluate these conditions across four task types--summarization, planning, explanation, and coding--using three LLM systems: ChatGPT, Claude, and Grok. Each output is scored with a unified rubric covering task completion, correctness, compliance, and clarity. Checklist-improved prompts achieved the highest mean rubric score, 7.50 out of 8, compared with 5.67 for raw prompts and 6.67 for clarifying-question prompts. Checklist prompts also produced the best quality-effort tradeoff, using fewer average tokens than both raw and clarifying prompts. These results suggest that a simple prompt checklist can improve LLM responses while reducing unnecessary interaction.

Saurav Ghosh, Gabriella Polach, Abdou Sow

Washington University in St. Louis Saint Louis, Missouri, USA

LLM EvaluationStructured PromptingLanguage Model Generation EvaluationLLM PromptingPrompt Engineering
AskPDF
May 19, 2026

LP-Eval: Rubric and Dataset for Measuring the Quality of Legal Proposition Generation

Legal proposition generation is central to legal reasoning and doctrinal scholarship, yet remain under-examined in Legal NLP. This paper investigates the automatic generation and evaluation of legal propositions from decisions of the Court of Justice of the European Union using large language models (LLMs). We introduce LP-Eval, a three-step evaluation rubric co-designed with legal experts that decomposes legal proposition quality into formal validity and substantive dimensions. Using this rubric, we release a dataset of two experts' annotations for 100 LLM-generated legal propositions. Our results show that LLMs can generate predominantly well-formed and high-quality propositions, while expert evaluations reveal higher quality for propositions derived from well established cases than from recent ones. We further examine LLMs as evaluators and find that rubric-guided LLM judgments align more closely with expert assessments than direct overall scoring, but remain insensitive to finer-grained distinctions captured by human experts.

Shanshan Xu, Johan Lindholm, Amogh Raina +2

University of Copenhagen · Umeå University

Legal NLPLLM EvaluationLLM-as-a-JudgeLanguage Model Generation EvaluationLegal ReasoningText Generation
AskPDF
May 19, 2026

Base Models Look Human To AI Detectors

As AI-generated text enters the real-world at scale, institutions increasingly use commercial AI-text detectors, especially in education and academic-integrity workflows. We report a surprising empirical finding about such systems: when evaluated by GPTZero and Pangram, generated text from base models is often judged overwhelmingly human, whereas text generated by their instruction-tuned counterparts is not. Building on this observation, we propose Humanization by Iterative Paraphrasing (HIP), a detector-agnostic pipeline that minimally fine-tunes a base model into a paraphraser and applies it iteratively. Compared with the baselines we test, HIP yields a stronger trade-off between semantic preservation and detector evasion on commercial detectors. Across Llama-3 and Qwen-3 families, spanning model sizes from 0.6B to 70B, HIP consistently improves detector human-likeness. Our findings suggest that current detectors are tracking artifacts of instruction tuning and local context more than any invariant notion of machine-generated text. This, in turn, calls for detector designs that model these factors more explicitly.

Yixuan Even Xu, Ziqian Zhong, Aditi Raghunathan +2

Carnegie Mellon University

AI-Generated Text DetectionLanguage Model Generation EvaluationInstruction TuningRobustness of Machine-Generated Text Detection
AskPDF
May 14, 2026

Dimension-Level Intent Fidelity Evaluation for Large Language Models: Evidence from Structured Prompt Ablation

Holistic evaluation scores capture overall output quality but do not distinguish whether a model reproduced the structural form of a user's request from whether it preserved the user's specific intent. We propose a dimension-level intent fidelity evaluation framework, applied here through a structured prompt ablation study across 2,880 outputs spanning three languages, three task domains, and six LLMs, that separately measures structural recovery and intent fidelity for each semantic dimension. This framework reveals a systematic structural-fidelity split: among Chinese-language outputs with complete paired scores, 25.7% received perfect holistic alignment scores (GA=5) while exhibiting measurable dimensional intent deficits; among English-language outputs, this proportion rose to 58.6%. Human evaluation confirmed that these split-zone outputs represent genuine quality deficits and that dimensional fidelity scores track human judgements more reliably than holistic scores do. A public-private decomposition of 2,520 ablation cells characterises when models successfully compensate for missing intent and when they fail, while proxy annotation distinguishes prior inferability from default recoverability. A weight-perturbation experiment shows that moderate misalignment is typically absorbed, whereas severe dimensional inversion is consistently harmful. These findings demonstrate that dimension-level intent fidelity evaluation is a necessary complement to holistic assessment when evaluating LLM outputs for user-specific tasks.

GAng Peng

Huizhou Lateni AI Technology Co., Ltd., Huizhou, China · Huizhou University, Huizhou, China

Multilingual Language Model EvaluationLLM EvaluationLanguage Model Generation EvaluationInstruction Following
AskPDF
May 14, 2026

NodeSyn: Socially Aligned Synthetic Data for AI Evaluation

Recent advancements in generative AI facilitate large-scale synthetic data generation for model evaluation. However, without targeted approaches, these datasets often lack the sociotechnical nuance required for sensitive domains. We introduce NodeSyn, a methodology that pairs a fine-tuned taxonomy generator (TaG) with credible published sources and extracted sensitive attributes to generate queries reflecting authentic societal contexts and nuanced harms. Evaluated across ten open-weight and proprietary LLMs in Medical Advice and Self-Harm using human-validated autoraters, NodeSyn ranked first on 8 of 10 models in each domain. NodeSyn datasets induced average failure rates 3.1 times higher than generic synthetic baselines, 2.4 times higher than structured synthetic benchmarks, and 1.6 times higher than authentic human-authored queries. Ablation studies confirm that, controlling for structural confounds, our granular Level 3 taxonomic expansion significantly drives these failure rates over Level 2 across 9 of 10 models (+29.2% relative lift in Medical Advice and +24.0% in Self-Harm). Variance attribution analysis reveals that failures are driven not only by what topics are queried (taxonomic nodes and lexical specificity), but also by their sociotechnical grounding (affected demographics, occupations, geographies, and querying personas). We embed this method into a research prototype, and open-source (https://github.com/google-research/nodesynth) the code and datasets to enable scalable, high-stakes model evaluation.

Qazi Mamunur Rashid, Xuan Yang, Zhengzhe Yang +5

Google Research, USA · Google, USA · Google DeepMind, USA

Language Model Generation EvaluationLanguage Model Safety EvaluationLLM AuditingSynthetic Data GenerationSynthetic Benchmark GenerationSynthetic Data Generation for Language Models+1
AskPDF
May 13, 2026

RTLC -- Research, Teach-to-Learn, Critique: A three-stage prompting paradigm inspired by the Feynman Learning Technique that lifts LLM-as-judge accuracy on JudgeBench with no fine-tuning

LLM-as-a-judge is now the default measurement instrument for open-ended generation, but on the public JudgeBench benchmark even strong instruction-tuned judges barely scrape past random on objective-correctness pairwise items. We introduce RTLC, a three-stage prompting recipe -- Research, Teach-to-Learn, Critique -- that promotes a single black-box LLM into an ensemble-of-thought judge with no fine-tuning, retrieval, or external tools. Stage 1 wraps the input in a fixed pedagogical scaffold porting the Feynman Learning Technique (study →\to teach →\to find gaps →\to simplify) into LLM prompting. Stage 2 draws N=10 independent candidate verdicts at temperature 0.4. Stage 3 acts as its own critic, cross-comparing the candidate set against the original question to emit one critiqued verdict at temperature 0. On JudgeBench-GPT (350 hard pairwise items), Claude 3.7 Sonnet's pairwise accuracy climbs from 64.6% (single-shot vanilla prompt) to 78.6% (RTLC critique-of-10) -- an absolute 14.0-percentage-point gain. RTLC also beats N=10 self-consistency majority voting (77.7%) and a zero-shot first candidate (74.0%). A clean three-step ablation attributes +9.4 pp to the Teach-to-Learn scaffold, +3.7 pp to N=10 marginalisation, and +0.9 pp to explicit critique. We discuss the cost-accuracy frontier (RTLC sits above self-consistency at every working point), the error-budget breakdown across the four JudgeBench categories (knowledge, reasoning, math, coding), and how RTLC composes orthogonally with post-hoc judge-score calibration, with the two interventions compounding multiplicatively in practice.

Andrea Morandi

Cisco Systems, Inc.

LLM EvaluationPairwise ComparisonLLM-as-a-JudgeLanguage Model Generation Evaluation
AskPDF
May 12, 2026

Predicting Disagreement with Human Raters in LLM-as-a-Judge Difficulty Assessment without Using Generation-Time Probability Signals

Automatic generation of educational materials using large language models (LLMs) is becoming increasingly common, but assigning difficulty levels to such materials still requires substantial human effort. LLM-as-a-Judge has therefore attracted attention, yet disagreement with human raters remains a major challenge. We propose a method for predicting which LLM-generated difficulty ratings are likely to disagree with human raters, so that such cases can be sent for re-rating. Unlike prior approaches, our method does not rely on generation-time probability signals, which must be collected during rating generation and are often difficult to compare across LLMs. Instead, exploiting the fact that difficulty is an ordinal scale, we use a separate embedding space, such as ModernBERT, and identify disagreement candidates based on the geometric consistency of the rating set. Experiments on English CEFR-based sentence difficulty assessment with GPT-OSS-120B and Qwen3-235B-A22B showed that the proposed method achieved higher AUC for predicting disagreement with human raters than probability-based baselines.

Yo Ehara

Tokyo Gakugei University

Annotator DisagreementLLM-as-a-JudgeLanguage Model Generation EvaluationEducational Assessment
AskPDF
May 12, 2026

DiffScore: Text Evaluation Beyond Autoregressive Likelihood

Autoregressive language models are widely used for text evaluation, however, their left-to-right factorization introduces positional bias, i.e., early tokens are scored with only leftward context, conflating architectural asymmetry with true text quality. We propose masked reconstruction as an alternative paradigm, where every token is scored using full bidirectional context. We introduce DiffScore, an evaluation framework built on Masked Large Diffusion Language Models. By measuring text recoverability across continuous masking rates, DiffScore eliminates positional bias and naturally establishes an evaluation hierarchy from local fluency to global coherence. We further provide diagnostic tools unavailable to autoregressive frameworks: multi-timestep quality profiles that decompose scores across masking rates, and bidirectional PMI decomposition that disentangles fluency from faithfulness. Experiments across ten benchmarks show that DiffScore consistently outperforms autoregressive baselines in both zero-shot and fine-tuned settings. The code is released at: https://github.com/wenlai-lavine/DiffScore.

Wen Lai, Yingli Shen, Dingnan Jin +4

1Ant Group · 2Tsinghua University · 3Technical University of Munich

LLM EvaluationLanguage Model Generation EvaluationAutomated EvaluationDiffusion Language ModelsDiffusion Language Model Benchmarking
AskPDF
May 11, 2026

VERDI: Single-Call Confidence Estimation for Verification-Based LLM Judges via Decomposed Inference

LLM-as-Judge systems are widely deployed for automated evaluation, yet practitioners lack reliable methods to know when a judge's verdict should be trusted. Token log-probabilities, the standard post-hoc confidence signal, are unavailable for many commercial LLMs and, even when accessible, saturate above 0.999 with structured JSON output. We introduce VERDI (VERification-Decomposed Inference), a method that extracts confidence from the reasoning trace a structured judge already produces, with no additional inference calls. VERDI decomposes each verification-style evaluation into sub-checks and derives three structural signals: Step-Verdict Alignment, Claim-Level Margin, and Evidence Grounding Score. We combine them with Platt-scaled logistic regression. On three public benchmarks, VERDI achieves AUROC 0.72-0.91 on GPT-4.1-mini and 0.66-0.80 on GPT-5.4-mini. On Qwen3.5-4B/9B/27B, where answer-token logprobs are anti-calibrated (higher confidence on errors, AUROC 0.32-0.49), VERDI achieves 0.56-0.70. We additionally validate on a production system with eight rubrics (AUROC 0.73-0.88 on factual rubrics), demonstrate cross-model transfer (AUROC 0.66-0.69), and show that a 33M-parameter NLI (Natural Language Inference) model provides a scalable alternative to regex extraction.

Jasmine Qi, Danylo Dantsev, Muyang Sun

Indeed Inc.

Post-Hoc CalibrationLLM-as-a-JudgeLLM Answer VerificationLanguage Model Generation EvaluationConfidence Estimation in Language ModelsLLM Reliability+1
AskPDF
May 11, 2026

LLARS: Enabling Domain Expert & Developer Collaboration for LLM Prompting, Generation and Evaluation

We demonstrate LLARS (LLM Assisted Research System), an open-source platform that bridges the gap between domain experts and developers for building LLM-based systems. It integrates three tightly connected modules into an end-to-end pipeline: Collaborative Prompt Engineering for real-time co-authoring with version control and instant LLM testing, Batch Generation for configurable output production across user-selected prompts ×\times models ×\times data with cost control, and Hybrid Evaluation where human and LLM evaluators jointly assess outputs through diverse assessment methods, with live agreement metrics and provenance analysis to identify the best model-prompt combination for a given use case. New prompts and models are automatically available for batch generation and completed batches can be turned into evaluation scenarios with a single click. Interviews with six domain experts and three developers in online counselling confirmed that LLARS feels intuitive, saves considerable time by keeping everything in one place and makes interdisciplinary collaboration seamless.

Philipp Steigerwald, Mara Stieler, Jennifer Burghardt +2

Technische Hochschule Nürnberg Georg Simon Ohm, Nürnberg, Germany · Faculty of Computer Science, Centre for Artificial Intelligence (KIZ) · Faculty of Social Sciences, Institute for E-Counselling

Language Model Generation EvaluationLLM PromptingHuman-in-the-Loop EvaluationPrompt Engineering
AskPDF
May 11, 2026

StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs

Multilingual studies of social bias in open-ended LLM generation remain limited: most existing benchmarks are English-centric, template-based, or restricted to recognizing pre-specified stereotypes. We introduce StereoTales, a multilingual dataset and evaluation pipeline for systematically studying the emergence of social bias in open-ended LLM generation. The dataset covers 10 languages and 79 socio-demographic attributes, and comprises over 650k stories generated by 23 recent LLMs, each annotated with the socio-demographic profile of the protagonist across 19 dimensions. From these, we apply statistical tests to identify more than 1{,}500 over-represented associations, which we then rate for harmfulness through both a panel of humans (N = 247) and the same LLMs. We report three main findings. (i) Every model we evaluate emits consequential harmful stereotypes in open-ended generation, regardless of size or capabilities, and these associations are largely shared across providers rather than isolated misbehaviors. (ii) Prompt language strongly shapes which stereotypes appear: rather than transferring as a shared set of biases, harmful associations adapt culturally to the prompt language and amplify bias against locally salient protected groups. (iii) Human and LLM harmfulness judgments are broadly aligned (Spearman ρ=0.62ρ=0.62), with disagreements concentrating on specific attribute classes rather than specific providers. To support further analyses, we release the evaluation code and the dataset, including model generations, attribute annotations, and harmfulness ratings.

Pierre Le Jeune, Étienne Duchesne, Weixuan Xiao +4

Giskard AI · École Normale Supérieure, INSERM, Paris

Open-Ended GenerationMultilingual Language Model EvaluationLanguage Model Generation EvaluationSocial Bias in Language ModelsCultural Bias in Language ModelsLanguage Model Bias Evaluation
AskPDF
May 11, 2026

Active Testing of Large Language Models via Approximate Neyman Allocation

Large language models (LLMs) require reliable evaluation from pre-training to test-time scaling, making evaluation a recurring rather than one-off cost. As model scales grow and target tasks increasingly demand expert annotators, both the compute and labeling costs needed for each evaluation rise rapidly. Active testing aims to alleviate this bottleneck by approximating the evaluation result from a small but informative subset of the evaluation pool. However, existing approaches primarily target classification and break down on generative tasks. We introduce a novel active testing algorithm tailored to generative tasks. Our method leverages semantic entropy from surrogate models to stratify the evaluation pool and then conducts approximate Neyman allocation based on signals extracted from these surrogates. Across multiple language and multimodal benchmarks and a range of surrogate-target model pairs, our method significantly improves on baselines and closely tracks Oracle-Neyman, delivering up to 28% MSE reduction over Uniform Sampling and an average of 22.9% budget savings.

Zeli Liu, Jiancheng Zhang, Cong Liu +1

University of California, Riverside

LLM EvaluationLanguage Model Generation EvaluationSemantic EntropyGenerative AI Evaluation
AskPDF
May 8, 2026

NARRA-Gym for Evaluating Interactive Narrative Agents

Interactive narrative tasks require LLMs to sustain a coherent, evolving story while adapting to a user over multiple turns. However, suitable benchmarks for this setting are limited: existing evaluations often focus on static prompts, isolated story generations, or post-hoc ratings, and therefore miss whether models can jointly manage story generation, long-context state and pacing, character simulation, empathic personalization, and story-grounded artifacts. We introduce NARRA-Gym, an executable evaluation environment that turns a sparse emotional seed into a complete interactive story episode and logs the full model-in-the-loop trajectory, including story construction, memory updates, planning, pacing interventions, and optional artifact synthesis. We evaluate nine frontier LLMs using a controlled LLM-as-judge sweep over eight benchmark personas and a human evaluation in which participants rate customized model outputs. Our results show substantial variation across models, personas, and evaluation dimensions: models that produce fluent stories can still fail on robustness, user experience, or resistance-sensitive personalization. These findings suggest that interactive narrative offers a useful benchmark for evaluating long-horizon, user-adaptive LLM behavior beyond isolated story quality.

Yue Huang, Yuchen Ma, Jiayi Ye +14

University of Notre Dame · 2LMU Munich · 11Munich Center for Machine Learning +8

Language Model Generation EvaluationInteractive StorytellingLLM Agent EvaluationLong-Horizon LLM AgentsNarrative Generation
AskPDF
May 7, 2026

Delulu: A Verified Multi-Lingual Benchmark for Code Hallucination Detection in Fill-in-the-Middle Tasks

Large Language Models for code generation frequently produce hallucinations in Fill-in-the-Middle (FIM) tasks -- plausible but incorrect completions such as invented API methods, invalid parameters, undefined variables, or non-existent imports. These failures pass superficial review yet introduce runtime errors. We introduce Delulu, a verified multi-lingual benchmark of 1,951 FIM samples across 7 languages and 4 hallucination types. Samples are curated through an adversarial pipeline: a frontier LLM generates plausible hallucinations, four diverse judge models evaluate them, embedding-based clustering mines progressively harder examples, self-contained Docker containers verify that golden completions compile while hallucinated variants produce the expected runtime error, and a final human-expert review removes any remaining biased or trivially decidable samples. We evaluate 11 open-weight FIM models from five families spanning 0.5B-32B parameters: a six-point Qwen2.5-Coder scaling slate, plus a cross-family slate (CodeLlama, DeepSeek-Coder-V2, StarCoder2). The strongest model reaches only 84.5% pass@1, no family exceeds 0.77 Edit Similarity, and every family produces hallucination-aligned completions on a non-trivial share of samples, confirming that the difficulty exposed by Delulu is task-intrinsic rather than family-specific. We release the benchmark, containers, and evaluation framework at https://github.com/microsoft/delulu.

Mahdi Erfanian, Nelson Daniel Troncoso, Aashna Garg +4

University of Illinois Chicago · Microsoft

Language Model Error DetectionLanguage Model Generation EvaluationHallucination in Language ModelsCode GenerationLLM ReliabilityLLM Hallucination Detection+2
AskPDF
May 7, 2026

ReasonLab: A Controlled and Auditable Evaluation of Prompting Techniques for Multiple-Choice QA

Probing the capabilities of Large Language Models (LLMs) and building robust solutions for Multiple-Choice Question Answering (MCQA) remain central challenges in natural language understanding. Furthermore, the rapid proliferation of LLMs has created the implicit assumption that more sophisticated prompting techniques yield better performance. Several studies claim such gains, but report them under differing models, prompt wordings and answer-extraction rules, so the gains cannot be attributed to the technique alone. We address this gap with ReasonLab, an evaluation framework in which the prompting technique is a first-class experimental variable alongside the model and the dataset, and which retains every generation for inspection. Using ReasonLab we conduct a controlled study of 8 prompting techniques across 10 MCQA datasets, 27 model configurations and 480,927 evaluations at temperature 0. We find that the prompting technique is a minor determinant of accuracy: on configurations without a reasoning budget the reasoning triggers improve on direct prompting by only 3.92 to 4.69 pp and are indistinguishable from one another, and on configurations with reasoning enabled no technique differs by more than 0.51 pp. Self-Generate is the only technique with a consistent effect, a reduction of 2.95 pp. We further investigate three phenomena: (1) the comparison of models on a common set of datasets, where model size does not predict accuracy, (2) the trade-offs across thinking budgets, where enabling reasoning is worth up to 12.74 pp whereas an eightfold budget increase adds only 0.48 to 2.10 pp, and (3) the variation in dataset difficulty, with 60% of benchmarks below 70% accuracy and a 43.9 pp spread from easiest to hardest. These results suggest that, for MCQA, the prompting technique is a minor lever compared with enabling model reasoning, and that substantial headroom remains.

Inder Preet, Shuxin Lin, Dhaval Patel

IBM, Dublin · IBM, New York

LLM EvaluationLanguage Model Generation EvaluationMultiple-Choice Question AnsweringLLM PromptingLarge Reasoning Models
AskPDF
May 7, 2026

Beyond Fixed Benchmarks and Worst-Case Attacks: Dynamic Boundary Evaluation for Language Models

Evaluating large language models (LLMs) today rests on fixed benchmarks that apply the same set of items to any model, producing ceiling and floor effects that mask capability gaps. We argue that the most informative evaluation signal lies at the boundary, where the per-prompt pass probability is near 0.50.5 under random-sampling decoding strategy, and propose Dynamic Boundary Evaluation (DBE), which actively locates each model's boundary and places it on a globally comparable difficulty scale. DBE delivers three artifacts: (i) a calibrated item bank covering safety, capability, and truthfulness, with per-item difficulty labels validated across 99 reference LLMs; (ii) Skill-Guided Boundary Search (SGBS), a search algorithm that finds boundary items for a given target LLM using only API-level query access; and (iii) an evaluation protocol that places a new LLM on a unified ability scale and grows the evaluation set adaptively when the target falls outside the bank's coverage. We instantiate DBE on four categories spanning safety (harmful request refusal, over-refusal), capability (constrained instruction following), and truthfulness (multi-turn sycophancy resistance). The resulting evaluation covers a broader model spectrum without saturation while remaining compatible with existing datasets.

Haoxiang Wang, Da Yu, Huishuai Zhang

Peking University · Google Research · Wangxuan Institute of Computer Technology, Peking University State Key Laboratory of General Artificial Intelligence

LLM EvaluationLanguage Model Generation EvaluationLanguage Model Safety Evaluation
AskPDF
May 6, 2026

Automatically Finding and Validating Unexpected Side-Effects of Interventions on Language Models

We present an automated, contrastive evaluation pipeline for auditing the behavioral impact of interventions on large language models. Given a base model M1M_1 and an intervention model M2M_2, our method compares their free-form, multi-token generations across aligned prompt contexts and produces human-readable, statistically validated natural-language hypotheses describing how the models differ, along with recurring themes that summarize patterns across validated hypotheses. We evaluate the approach in synthetic setting by injecting known behavioral changes and showing that the pipeline reliably recovers them. We then apply it to three real-world interventions, reasoning distillation, knowledge editing and unlearning, demonstrating that the method surfaces both intended and unexpected behavioral shifts, distinguishes large from subtle interventions, and does not hallucinate differences when effects are absent or misaligned with the prompt bank. Overall, the pipeline provides a statistically grounded and interpretable tool for post-hoc auditing of intervention-induced changes in model behavior.

Quintin Pope, Ajay Hayagreeve Balaji, Jacques Thibodeau +1

Oregon State University · Independent Researcher

Language Model Generation EvaluationLLM AuditingCausal Interventions in Language Models
AskPDF
May 6, 2026

StoryAlign: Evaluating and Training Reward Models for Story Generation

Story generation aims to automatically produce coherent, structured, and engaging narratives. Although large language models (LLMs) have significantly advanced text generation, stories generated by LLMs still diverge from human-authored works regarding complex narrative structure and human-aligned preferences. A key reason is the absence of effective modeling of human story preferences, which are inherently subjective and under-explored. In this work, we systematically evaluate the modeling of human story preferences and introduce StoryRMB, the first benchmark for assessing reward models on story preferences. StoryRMB contains 1,1331,133 high-quality, human-verified instances, each consisting of a prompt, one chosen story, and three rejected stories. We find existing reward models struggle to select human-preferred stories, with the best model achieving only 66.3%66.3\% accuracy. To address this limitation, we construct roughly 100,000100,000 high-quality story preference pairs across diverse domains and develop StoryReward, an advanced reward model for story preference trained on this dataset. StoryReward achieves state-of-the-art (SoTA) performance on StoryRMB, outperforming much larger models. We also adopt StoryReward in downstream test-time scaling applications for best-of-n (BoN) story selection and find that it generally chooses stories better aligned with human preferences. We will release our dataset, model, and code to facilitate future research. Related code and data are available at https://github.com/THU-KEG/StoryReward.

Haotian Xia, Hao Peng, Yunjia Qi +4

Department of Computer Science and Technology, Tsinghua University

Reward ModelingLanguage Model Generation EvaluationHuman Preference ModelingText GenerationReward Model EvaluationNarrative Generation
AskPDF
May 6, 2026

RaguTeam at SemEval-2026 Task 8: Meno and Friends in a Judge-Orchestrated LLM Ensemble for Faithful Multi-Turn Response Generation

We present our winning system for TaskB (generation with reference passages) in SemEval-2026 Task8: MTRAGEval. Our method is a heterogeneous ensemble of seven LLMs with two prompting variants, where a GPT-4o-mini judge selects the best candidate per instance. We ranked 1st out of 26 teams, achieving a conditioned harmonic mean of 0.7827 and outperforming the strongest baseline (gpt-oss-120b, 0.6390). Ablations show that diversity in model families, scales, and prompting strategies is essential, with the ensemble consistently beating any single model. We also introduce Meno-Lite-0.1, a 7B domain-adapted model with a strong cost--performance trade-off, and analyse MTRAGEval, highlighting annotation limitations and directions for improvement. Our code is publicly available: https://github.com/RaguTeam/ragu_mtrag_semeval

Ivan Bondarenko, Roman Derunets, Oleg Sedukhin +3

Novosibirsk State University

Language Model Generation EvaluationRAG EvaluationLanguage Model Ensembles
AskPDF
May 6, 2026

DoGMaTiQ: Automated Generation of Question-and-Answer Nuggets for Report Evaluation

Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems. Core to many evaluation frameworks is the use of atomic facts, or nuggets, to assess a report's coverage of query-relevant information attested in the underlying collection. While nuggets have traditionally been represented as short statements, recent work has used question-answer (QA) representations, enabling fine-grained evaluations that decouple the information need (i.e. the question) from the potentially diverse content that satisfies it (i.e. its answers). A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs. This challenge is acute in cross-lingual settings, where information is found in multilingual source documents. Accordingly, we introduce DoGMaTiQ, a pipeline for generating high-quality QA-based nugget sets in three stages: (1) document-grounded nugget generation, (2) paraphrase clustering, and (3) nugget subselection based on principled quality criteria. We integrate DoGMaTiQ nuggets with AutoArgue -- a recent nugget-based evaluation framework -- to enable fully automatic evaluation of generated reports. We conduct extensive experiments on two cross-lingual TREC shared tasks, NeuCLIR and RAGTIME, showing strong rank correlations with both human-in-the-loop and fully manual judgments. Finally, detailed analysis of our pipeline reveals that a strong LLM nugget generator is key, and that the system rankings induced by DoGMaTiQ are robust to outlier systems. We facilitate future research in report evaluation by publicly releasing our code and artifacts at https://github.com/manestay/dogmatiq.

Bryan Li, William Walden, Yu Hou +6

Google Inc. · University of Pennsylvania · Johns Hopkins University +3

Language Model Generation EvaluationRAG EvaluationAutomated EvaluationIR Evaluation
AskPDF
May 5, 2026

EvoLM: Self-Evolving Language Models through Co-Evolved Discriminative Rubrics

Language models encode substantial evaluative knowledge from pretraining, yet current post-training methods rely on external supervision (human annotations, proprietary models, or scalar reward models) to produce reward signals. Each imposes a ceiling. Human judgment cannot supervise capabilities beyond its own, proprietary APIs create dependencies, and verifiable rewards cover only domains with ground-truth answers. Self-improvement from a model's own evaluative capacity is a reward source that scales with the model itself, yet remains largely untapped by current methods. We introduce EVOLM, a post-training method that structures this capacity into explicit discriminative rubrics and uses them as training signal. EVOLM trains two capabilities within a single language model in alternation: (1) a rubric generator producing instance-specific evaluation criteria optimized for discriminative utility, which maximizes a small frozen judge's ability to distinguish preferred from dispreferred responses; and (2) a policy trained using those rubric-conditioned scores as reward. All preference signals are constructed from the policy's own outputs via temporal contrast with earlier checkpoints, requiring no human annotation or external supervision. EVOLM trains a Qwen3-8B model to generate rubrics that outperform GPT-4.1 on RewardBench-2 by 25.7%. The co-trained policy achieves 69.3% average on the OLMo3-Adapt suite, outperforming policies trained with GPT-4.1 prompted rubrics by 3.9% and with the state-of-the-art 8B reward model SkyWork-RM by 16%. Overall, EVOLM demonstrates that structuring a model's evaluative capacity into co-evolving discriminative rubrics enables self-improvement without external supervision.

Shuyue Stella Li, Rui Xin, Teng Xiao +8

University of Washington · 2Allen Institute for AI · University of Pennsylvania

RL for Language ModelsLLM EvaluationLanguage Model Generation EvaluationSelf-Training for Language ModelsLanguage Model Self-AssessmentLanguage Model Post-Training+1
AskPDF
May 4, 2026

Beating the Style Detector: Three Hours of Agentic Research on the AI-Text Arms Race

Reproducing an empirical NLP study used to take weeks. Given the released data and a modern agentic-research harness, we redo every experiment of a recent ACL,2026 study on personal-style post-editing of LLM drafts -- and add three new ones -- with the human investigator acting only as a reviewer-in-the-loop. We reproduce all seven preregistered hypotheses and recover the paper's headline correlation between perceived self-similarity and embedding-measured self-similarity to three decimal places (r=+0.244r{=}{+}0.244, p<10−8p{<}10^{-8}, n=648n{=}648). Under a leakage-free held-out protocol, GPT-5.5 and Claude,Opus,4.7 close 7171--75 %75\,\% of the style gap to the same-author ceiling on 324324 paired tasks, against 24 %24\,\% for the human post-edit, and beat the human post-edit on \sim$$80\,\% of tasks. We then frame the same data as an AI-text detection arms race. A leave-authors-out linear SVM on LUAR-MUD embeddings reaches AUC 0.930.93--1.001.00 across approaches; six diagnostics show that GPT-5.5 detection is mostly a length confound while Opus detection is a genuine stylistic signature. Given T=20T{=}20 feedback iterations against the frozen detector, an Opus agent flips two of five held-out test mimics to the human half-space and shrinks every margin by an order of magnitude. With moderate effort against a known detector, a frontier LLM can already efficiently lower its own AI-detection probability. All code, 648648 mimic drafts, trained detectors, diagnostics, and adversarial trajectories are released.

Andreas Maier, Moritz Zaiss, Siming Bayer

Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany · Institute of Neuroradiology, University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany

AI-Generated Text DetectionLanguage Model Generation EvaluationText Style TransferRobustness of Machine-Generated Text DetectionText GenerationAdversarial Text Attacks
AskPDF