mlsanityData to 14 Oct 2026

Self-critiquing models for assisting human evaluators

arXiv 2206.05802 ↗First posted 12 Jun 2022cs.CL cs.LGcited by 380

Not on the map. Why?

Cited by 99 in LLM reasoning, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We fine-tune large language models to write natural language critiques (natural language critical comments) using behavioral cloning. On a topic-based summarization task, critiques written by our models help humans find flaws in summaries that they would have otherwise missed. Our models help find naturally occurring flaws in both model and human written summaries, and intentional flaws in summaries written by humans to be deliberately misleading. We study scaling properties of critiquing with both topic-based summarization and synthetic tasks. Larger models write more helpful critiques, and on most tasks, are better at self-critiquing, despite having harder-to-critique outputs. Larger models can also integrate their own self-critiques as feedback, refining their own summaries into better ones. Finally, we motivate and introduce a framework for comparing critiquing ability to generation and discrimination ability. Our measurements suggest that even large models may still have relevant knowledge they cannot or do not articulate as critiques. These results are a proof of concept for using AI-assisted human feedback to scale the supervision of machine learning systems to tasks that are difficult for humans to evaluate directly. We release our training datasets, as well as samples from our critique assistance experiments.

Key facts. Cited by 380 papers in the map (data through 14 Oct 2026). Builds on 25 of its 46 references that are in the map. First posted on arXiv 12 Jun 2022.

Builds on

25 of 46 references.

References · 25

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2020Scaling Laws for Neural Language Modelscited by 6,586
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  5. 2016Concrete Problems in AI Safetycited by 1,642
  6. 2022Competition-Level Code Generation with AlphaCodecited by 1,462
  7. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,375
  8. 2022Red Teaming Language Models with Language Modelscited by 1,026
  9. 2017RACE: Large-scale ReAding Comprehension Dataset From Examinationscited by 861
  10. 2019GLTR: Statistical Detection and Visualization of Generated Textcited by 496
  11. 2018Scalable agent alignment via reward modeling: a research directioncited by 447
  12. 2018AI safety via debatecited by 400
13 more
  1. 2021Recursively Summarizing Books with Human Feedbackcited by 290
  2. 2022WANLI: Worker and AI Collaboration for Natural Language Inference Dataset Creationcited by 202
  3. 2018Categorizing Variants of Goodhart's Lawcited by 149
  4. 2018Supervising strong learners by amplifying weak expertscited by 145
  5. 2018A Dataset of Peer Reviews (PeerRead): Collection, Insights and NLP Applicationscited by 136
  6. 2020Reward-rational (implicit) choice: A unifying formalism for reward learningcited by 120
  7. 2020Generating Fact Checking Briefscited by 50
  8. 2020Speak to your Parser: Interactive Text-to-SQL with Natural Language Feedbackcited by 44
  9. 2021Learning to Give Checkable Answers with Prover-Verifier Gamescited by 32
  10. 2019Finding Generalizable Evidence by Learning to Convince Q&A Modelscited by 25
  11. 2022Single-Turn Debate Does Not Help Humans Answer Hard Reading-Comprehension Questionscited by 23
  12. 2018Guide Me: Interacting with Deep Networkscited by 16
  13. 2021Cut the CARP: Fishing for zero-shot story evaluationcited by 14

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 15; 2023: 86; 2024: 130; 2025: 92; 2026: 57.

380 papers.

Sort:

By area:

  1. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  2. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  3. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  4. 2023CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Societycited by 1,395
  5. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,302
  6. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Promptingcited by 1,122
  7. 2023Towards Understanding Sycophancy in Language Modelscited by 1,040
  8. 2023Teaching Large Language Models to Self-Debugcited by 988
  9. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 836
  10. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 783
  11. 2023Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedbackcited by 759
  12. 2023AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedbackcited by 748
38 more
  1. 2024LLM Evaluators Recognize and Favor Their Own Generationscited by 686
  2. 2024Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Modelscited by 557
  3. 2022Improving alignment of dialogue agents via targeted human judgementscited by 531
  4. 2024Generative Verifiers: Reward Modeling as Next-Token Predictioncited by 480
  5. 2023Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervisioncited by 467
  6. 2023Cognitive Architectures for Language Agentscited by 429
  7. 2024Training Language Models to Self-Correct via Reinforcement Learningcited by 426
  8. 2023UltraFeedback: Boosting Language Models with Scaled AI Feedbackcited by 406
  9. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 384
  10. 2023Language Models can Solve Computer Taskscited by 372
  11. 2022News Summarization and Evaluation in the Era of GPT-3cited by 327
  12. 2024JudgeBench: A Benchmark for Evaluating LLM-based Judgescited by 318
  13. 2023MINT: Evaluating LLMs in Multi-turn Interaction with Tools and Language Feedbackcited by 297
  14. 2024ProcessBench: Identifying Process Errors in Mathematical Reasoningcited by 292
  15. 2024A Survey on Knowledge Distillation of Large Language Modelscited by 284
  16. 2022Generating Sequences by Learning to Self-Correctcited by 277
  17. 2024Debating with More Persuasive LLMs Leads to More Truthful Answerscited by 266
  18. 2023A Survey of Reinforcement Learning from Human Feedbackcited by 264
  19. 2022The Alignment Problem from a Deep Learning Perspectivecited by 258
  20. 2024OpenCodeInterpreter: Integrating Code Generation with Execution and Refinementcited by 247
  21. 2024When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMscited by 240
  22. 2023Automatically Correcting Large Language Models: Surveying the landscape of diverse self-correction strategiescited by 238
  23. 2023Reward Model Ensembles Help Mitigate Overoptimizationcited by 231
  24. 2023AI Control: Improving Safety Despite Intentional Subversioncited by 229
  25. 2022Measuring Progress on Scalable Oversight for Large Language Modelscited by 225
  26. 2023REFINER: Reasoning Feedback on Intermediate Representationscited by 205
  27. 2024Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judgecited by 204
  28. 2022BlenderBot 3: a deployed conversational agent that continually learns to responsibly engagecited by 198
  29. 2023Simple synthetic data reduces sycophancy in large language modelscited by 187
  30. 2022Language models show human-like content effects on reasoning taskscited by 173
  31. 2023Generative Judge for Evaluating Alignmentcited by 169
  32. 2025Less-to-More Generalization: Unlocking More Controllability by In-Context Generationcited by 154
  33. 2023Self-Alignment with Instruction Backtranslationcited by 151
  34. 2023Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchillacited by 149
  35. 2024LLM Critics Help Catch LLM Bugscited by 147
  36. 2022Large Language Models Meet NL2Code: A Surveycited by 138
  37. 2025LLM Post-Training: A Deep Dive into Reasoning Large Language Modelscited by 135
  38. 2023Training Language Models with Language Feedback at Scalecited by 128

50 most cited of 380 All 380 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2022Training Language Models with Language Feedback
  2. 2020Learning to summarize from human feedback
  3. 2025Self-Critique and Refinement for Faithful Natural Language Explanations
  4. 2022On Improving Summarization Factual Consistency from Natural Language Feedback
  5. 2023Large Language Models are Not Yet Human-Level Evaluators for Abstractive Summarization
  6. 2023Human Feedback is not Gold Standard
  7. 2025Scalable Oversight for Superhuman AI via Recursive Self-Critiquing
  8. 2025Learning from Self Critique and Refinement for Faithful LLM Summarization