mlsanityData to 14 Oct 2026

Lessons from the Trenches on Reproducible Evaluation of Language Models

arXiv 2405.14782 ↗First posted 23 May 2024cs.CLcited by 179

Not on the map. Why?

Abstract (arXiv)

Reliable evaluation of language models (LMs) remains an open challenge. Re- searchers and engineers face methodological issues such as the sensitivity of models to evaluation setup, difficulty of proper comparisons across methods, and the lack of reproducibility and transparency. Evaluation difficulties are exacer- bated by the fracturing and siloing of information about conventions and common practices. In this paper we draw on three years of experience in evaluating large lan- guage models (LMs) as developers of the popular Language Model Evaluation Harness (lm-eval) (Gao et al., 2023) framework to provide guidance and lessons for the field moving forward. We document a variety of challenges faced by prac- titioners and provide concrete instances where these challenges or the absence of best practices have come into effect. We make recommendations to the field for improving evaluation rigor and confidence, and attempt to codify much of the tacit or folk knowledge surrounding LM evaluation, for a solid ground to move forward.

Key facts. Cited by 179 papers in the map (data through 14 Oct 2026). Builds on 96 of its 149 references that are in the map. First posted on arXiv 23 May 2024.

Builds on

96 of 149 references.

References · 96

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  5. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  6. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  7. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,326
  8. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  9. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  10. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  11. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
  12. 2016Pointer Sentinel Mixture Modelscited by 3,096
38 more
  1. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,975
  2. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
  3. 2019The Curious Case of Neural Text Degenerationcited by 2,916
  4. 2019PIQA: Reasoning about Physical Commonsense in Natural Languagecited by 2,883
  5. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  6. 2022Emergent Abilities of Large Language Modelscited by 2,512
  7. 2021Efficiently Modeling Long Sequences with Structured State Spacescited by 2,465
  8. 2020The Pile: An 800GB Dataset of Diverse Text for Language Modelingcited by 2,411
  9. 2023G-Eval: NLG Evaluation using GPT-4 with Better Human Alignmentcited by 2,276
  10. 2019BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questionscited by 2,183
  11. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,164
  12. 2023Pythia: A Suite for Analyzing Large Language Models Across Training and Scalingcited by 1,786
  13. 2019SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systemscited by 1,729
  14. 2019Do ImageNet Classifiers Generalize to ImageNet?cited by 1,515
  15. 2021Multitask Prompted Training Enables Zero-Shot Task Generalizationcited by 1,470
  16. 2022Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?cited by 1,434
  17. 2022Solving Quantitative Reasoning Problems with Language Modelscited by 1,392
  18. 2023Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generationcited by 1,391
  19. 2023PaLM 2 Technical Reportcited by 1,162
  20. 2021Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivitycited by 1,135
  21. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  22. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,055
  23. 2016The LAMBADA dataset: Word prediction requiring a broad discourse contextcited by 992
  24. 2019DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphscited by 903
  25. 2023RWKV: Reinventing RNNs for the Transformer Eracited by 827
  26. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 817
  27. 2023The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Onlycited by 763
  28. 2022GPT-NeoX-20B: An Open-Source Autoregressive Language Modelcited by 713
  29. 2020CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Modelscited by 710
  30. 2021BBQ: A Hand-Built Bias Benchmark for Question Answeringcited by 701
  31. 2020Long Range Arena: A Benchmark for Efficient Transformerscited by 700
  32. 2018Gender Bias in Coreference Resolution: Evaluation and Debiasing Methodscited by 659
  33. 2023Retentive Network: A Successor to Transformer for Large Language Modelscited by 613
  34. 2023Prometheus: Inducing Fine-grained Evaluation Capability in Language Modelscited by 575
  35. 2021Datasets: A Community Library for Natural Language Processingcited by 530
  36. 2023Are Emergent Abilities of Large Language Models a Mirage?cited by 523
  37. 2023"Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Modelscited by 489
  38. 2023Hyena Hierarchy: Towards Larger Convolutional Language Modelscited by 393

50 most cited of 96 All 96 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 24; 2025: 67; 2026: 88.

179 papers.

Sort:

By area:

  1. 2025s1: Simple test-time scalingcited by 1,341
  2. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
  3. 2025Kimi K2: Open Agentic Intelligencecited by 567
  4. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 415
  5. 2024OLMoE: Open Mixture-of-Experts Language Modelscited by 301
  6. 2024AI Agents That Mattercited by 219
  7. 2025Kimi Linear: An Expressive, Efficient Attention Architecturecited by 196
  8. 2024RegMix: Data Mixture as Regression for Language Model Pre-trainingcited by 177
  9. 2024LAB-Bench: Measuring Capabilities of Language Models for Biology Researchcited by 156
  10. 2024CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmarkcited by 123
  11. 2025Measuring what Matters: Construct Validity in Large Language Model Benchmarkscited by 119
  12. 2024BetterBench: Assessing AI Benchmarks, Uncovering Issues, and Establishing Best Practicescited by 114
38 more
  1. 2024OLMES: A Standard for Language Model Evaluationscited by 106
  2. 2025Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivitycited by 101
  3. 2025A Sober Look at Progress in Language Model Reasoning: Pitfalls and Paths to Reproducibilitycited by 93
  4. 2024Non-Determinism of "Deterministic" LLM Settingscited by 92
  5. 2025Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluationcited by 92
  6. 2024A Systematic Survey and Critical Review on Evaluating Large Language Models: Challenges, Limitations, and Recommendationscited by 85
  7. 2025Path Planning for Masked Diffusion Model Samplingcited by 80
  8. 2024Pangea: A Fully Open Multilingual Multimodal LLM for 39 Languagescited by 64
  9. 2025Can We Trust AI Benchmarks? An Interdisciplinary Review of Current Issues in AI Evaluationcited by 60
  10. 2025MesaNet: Sequence Modeling by Locally Optimal Test-Time Trainingcited by 47
  11. 2024IrokoBench: A New Benchmark for African Languages in the Age of Large Language Modelscited by 43
  12. 2025PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runscited by 35
  13. 2024Reproducibility in Machine Learning-based Research: Overview, Barriers and Driverscited by 34
  14. 2025Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmarkcited by 31
  15. 2025Parallel Scaling Law for Language Modelscited by 30
  16. 2025Position: Don't Use the CLT in LLM Evals With Fewer Than a Few Hundred Datapointscited by 23
  17. 2025Localized Cultural Knowledge is Conserved and Controllable in Large Language Modelscited by 20
  18. 2025Common Corpus: The Largest Collection of Ethical Data for LLM Pre-Trainingcited by 20
  19. 2025PoLAR: Polar-Decomposed Low-Rank Adapter Representationcited by 20
  20. 2025LLM-Safety Evaluations Lack Robustnesscited by 18
  21. 2025A Comedy of Estimators: On KL Regularization in RL Training of LLMscited by 18
  22. 2026On Randomness in Agentic Evalscited by 17
  23. 2025Don't Pass@k: A Bayesian Framework for Large Language Model Evaluationcited by 15
  24. 2026Tiny Aya: Bridging Scale and Multilingual Depthcited by 15
  25. 2026SafetyRepro: Configuration-Conditional Rank Instability on Alignment Benchmarkscited by 15
  26. 2025Improving Automatic Evaluation of Large Language Models (LLMs) in Biomedical Relation Extraction via LLMs-as-the-Judgecited by 13
  27. 2025Language Model Behavioral Phases are Consistent Across Architecture, Training Data, and Scalecited by 13
  28. 2025KurTail : Kurtosis-based LLM Quantizationcited by 12
  29. 2024MINERS: Multilingual Language Models as Semantic Retrieverscited by 11
  30. 2024The Responsible Foundation Model Development Cheatsheet: A Review of Tools & Resourcescited by 11
  31. 2025AFRIDOC-MT: Document-level MT Corpus for African Languagescited by 11
  32. 2025Who Benchmarks the Benchmarks? Towards Comprehensive Evaluation of Commonsense Reasoning Benchmarkscited by 11
  33. 2026Benchmarks Saturate When The Model Gets Smarter Than The Judgecited by 11
  34. 2026The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?cited by 10
  35. 2026AdaPonderLM: Gated Pondering Language Models with Token-Wise Adaptive Depthcited by 10
  36. 2024Leaving the barn door open for Clever Hans: Simple features predict LLM benchmark answerscited by 9
  37. 2025ReasonBENCH: Benchmarking the (In)Stability of LLM Reasoningcited by 9
  38. 2025Lugha-Llama: Adapting Large Language Models for African Languagescited by 8

50 most cited of 179 All 179 (JSON)