Language Model Generation Evaluation

Follow this topic Ask about this topic

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

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

Faithfulness Metrics Don't Measure Faithfulness: A Meta-Evaluation with Ground Truth
May 24, 2026Yoav Gur-Arieh, Ana Marasović, Mor GevaCoT Faithfulness LLM Evaluation
Tel Aviv University · University of Utah
Repeated Sequences Reveal Gaps between Large Language Models and Natural Language
May 24, 2026Kumiko Tanaka-IshiiLLM Evaluation Language Model Generation Evaluation
Waseda University, Japan.
Benchmarking the Personalization Capabilities of Large Language Models
May 23, 2026Ashutosh Srivastava, Siddharth Yedlapati, Vinay Aggarwal +4Language Model Generation Evaluation Personalized Language Models
Adobe Media & Data Science Research
NLG Evaluation: Past, Present, Future
May 22, 2026Ehud ReiterLLM Evaluation Language Model Generation Evaluation
Dept of Computing Science University of Aberdeen Aberdeen, UK
How Human-Like Are Large Language Models? A Register-Aware Linguistic Evaluation Framework
May 22, 2026Björn Nieth, Marianna Gracheva, Michaela Mahlberg +2LLM Evaluation Language Model Generation Evaluation
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
AMEL: Accumulated Message Effects on LLM Judgments
May 21, 2026Sid-Ali TemkitLLM Evaluation Language Model Generation Evaluation
RankJudge: A Multi-Turn LLM-as-a-Judge Synthetic Benchmark Generator
May 20, 2026Zhenwei Tang, Zhaoyan Liu, Rasa Hosseinzadeh +3LLM Evaluation Pairwise Comparison
Layer 6 AI · University of Toronto
APM: Evaluating Style Personalization in LLMs with Arbitrary Preference Mappings
May 20, 2026Philipp Spohn, Leander Girrbach, Zeynep AkataLLM Evaluation LLM Personalization
Technical University of Munich, Helmholtz Munich · Munich Center for Machine Learning (MCML)
Less Back-and-Forth: A Comparative Study of Structured Prompting
May 19, 2026Saurav Ghosh, Gabriella Polach, Abdou SowLLM Evaluation Structured Prompting
Washington University in St. Louis Saint Louis, Missouri, USA
LP-Eval: Rubric and Dataset for Measuring the Quality of Legal Proposition Generation
May 19, 2026Shanshan Xu, Johan Lindholm, Amogh Raina +2Legal NLP LLM Evaluation
University of Copenhagen · Umeå University
Base Models Look Human To AI Detectors
May 19, 2026Yixuan Even Xu, Ziqian Zhong, Aditi Raghunathan +2AI-Generated Text Detection Language Model Generation Evaluation
Carnegie Mellon University
Dimension-Level Intent Fidelity Evaluation for Large Language Models: Evidence from Structured Prompt Ablation
May 14, 2026GAng PengMultilingual Language Model Evaluation LLM Evaluation
Huizhou Lateni AI Technology Co., Ltd., Huizhou, China · Huizhou University, Huizhou, China
NodeSyn: Socially Aligned Synthetic Data for AI Evaluation
May 14, 2026Qazi Mamunur Rashid, Xuan Yang, Zhengzhe Yang +5Language Model Generation Evaluation Language Model Safety Evaluation
Google Research, USA · Google, USA · Google DeepMind, USA
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
May 13, 2026Andrea MorandiLLM Evaluation Pairwise Comparison
Cisco Systems, Inc.
Predicting Disagreement with Human Raters in LLM-as-a-Judge Difficulty Assessment without Using Generation-Time Probability Signals
May 12, 2026Yo EharaAnnotator Disagreement LLM-as-a-Judge
Tokyo Gakugei University
DiffScore: Text Evaluation Beyond Autoregressive Likelihood
May 12, 2026Wen Lai, Yingli Shen, Dingnan Jin +4LLM Evaluation Language Model Generation Evaluation
1Ant Group · 2Tsinghua University · 3Technical University of Munich
VERDI: Single-Call Confidence Estimation for Verification-Based LLM Judges via Decomposed Inference
May 11, 2026Jasmine Qi, Danylo Dantsev, Muyang SunPost-Hoc Calibration LLM-as-a-Judge
Indeed Inc.
LLARS: Enabling Domain Expert & Developer Collaboration for LLM Prompting, Generation and Evaluation
May 11, 2026Philipp Steigerwald, Mara Stieler, Jennifer Burghardt +2Language Model Generation Evaluation LLM Prompting
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
StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs
May 11, 2026Pierre Le Jeune, Étienne Duchesne, Weixuan Xiao +4Open-Ended Generation Multilingual Language Model Evaluation
Giskard AI · École Normale Supérieure, INSERM, Paris
Active Testing of Large Language Models via Approximate Neyman Allocation
May 11, 2026Zeli Liu, Jiancheng Zhang, Cong Liu +1LLM Evaluation Language Model Generation Evaluation
University of California, Riverside
NARRA-Gym for Evaluating Interactive Narrative Agents
May 8, 2026Yue Huang, Yuchen Ma, Jiayi Ye +14Language Model Generation Evaluation Interactive Storytelling
University of Notre Dame · 2LMU Munich · 11Munich Center for Machine Learning +8
Delulu: A Verified Multi-Lingual Benchmark for Code Hallucination Detection in Fill-in-the-Middle Tasks
May 7, 2026Mahdi Erfanian, Nelson Daniel Troncoso, Aashna Garg +4Language Model Error Detection Language Model Generation Evaluation
University of Illinois Chicago · Microsoft
ReasonLab: A Controlled and Auditable Evaluation of Prompting Techniques for Multiple-Choice QA
May 7, 2026Inder Preet, Shuxin Lin, Dhaval PatelLLM Evaluation Language Model Generation Evaluation
IBM, Dublin · IBM, New York
Beyond Fixed Benchmarks and Worst-Case Attacks: Dynamic Boundary Evaluation for Language Models
May 7, 2026Haoxiang Wang, Da Yu, Huishuai ZhangLLM Evaluation Language Model Generation Evaluation
Peking University · Google Research · Wangxuan Institute of Computer Technology, Peking University State Key Laboratory of General Artificial Intelligence
Automatically Finding and Validating Unexpected Side-Effects of Interventions on Language Models
May 6, 2026Quintin Pope, Ajay Hayagreeve Balaji, Jacques Thibodeau +1Language Model Generation Evaluation LLM Auditing
Oregon State University · Independent Researcher
StoryAlign: Evaluating and Training Reward Models for Story Generation
May 6, 2026Haotian Xia, Hao Peng, Yunjia Qi +4Reward Modeling Language Model Generation Evaluation
Department of Computer Science and Technology, Tsinghua University
RaguTeam at SemEval-2026 Task 8: Meno and Friends in a Judge-Orchestrated LLM Ensemble for Faithful Multi-Turn Response Generation
May 6, 2026Ivan Bondarenko, Roman Derunets, Oleg Sedukhin +3Language Model Generation Evaluation RAG Evaluation
Novosibirsk State University
DoGMaTiQ: Automated Generation of Question-and-Answer Nuggets for Report Evaluation
May 6, 2026Bryan Li, William Walden, Yu Hou +6Language Model Generation Evaluation RAG Evaluation
Google Inc. · University of Pennsylvania · Johns Hopkins University +3
EvoLM: Self-Evolving Language Models through Co-Evolved Discriminative Rubrics
May 5, 2026Shuyue Stella Li, Rui Xin, Teng Xiao +8RL for Language Models LLM Evaluation
University of Washington · 2Allen Institute for AI · University of Pennsylvania
Beating the Style Detector: Three Hours of Agentic Research on the AI-Text Arms Race
May 4, 2026Andreas Maier, Moritz Zaiss, Siming BayerAI-Generated Text Detection Language Model Generation Evaluation
Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany · Institute of Neuroradiology, University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany

May 24, 2026J/K move · Enter open · S save

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

Yoav Gur-Arieh, Ana Marasović, Mor Geva

Tel Aviv University · University of Utah

CoT Faithfulness LLM Evaluation Language Model Generation Evaluation Faithfulness of Language Model Explanations LLM Reliability

Open paper AskPDF arXiv

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.