mlsanityData to 7 Oct 2026

How is ChatGPT's behavior changing over time?

arXiv 2307.09009 ↗First posted 18 Jul 2023cs.CL cs.AI cs.LGcited by 265

Not on the map. Why?

Cited by 23 in LLM agents and tool use, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

GPT-3.5 and GPT-4 are the two most widely used large language model (LLM) services. However, when and how these models are updated over time is opaque. Here, we evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on several diverse tasks: 1) math problems, 2) sensitive/dangerous questions, 3) opinion surveys, 4) multi-hop knowledge-intensive questions, 5) generating code, 6) US Medical License tests, and 7) visual reasoning. We find that the performance and behavior of both GPT-3.5 and GPT-4 can vary greatly over time. For example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy). This is partly explained by a drop in GPT-4's amenity to follow chain-of-thought prompting. Interestingly, GPT-3.5 was much better in June than in March in this task. GPT-4 became less willing to answer sensitive questions and opinion survey questions in June than in March. GPT-4 performed better at multi-hop questions in June than in March, while GPT-3.5's performance dropped on this task. Both GPT-4 and GPT-3.5 had more formatting mistakes in code generation in June than in March. We provide evidence that GPT-4's ability to follow user instructions has decreased over time, which is one common factor behind the many behavior drifts. Overall, our findings show that the behavior of the "same" LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.

Key facts. Cited by 265 papers in the map (data through 7 Oct 2026). Builds on 22 of its 51 references that are in the map. First posted on arXiv 18 Jul 2023.

Builds on

22 of 51 references.

References · 22

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  4. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,686
  5. 2021Program Synthesis with Large Language Modelscited by 3,509
  6. 2023Sparks of Artificial General Intelligence: Early experiments with GPT-4cited by 2,579
  7. 2020Extracting Training Data from Large Language Modelscited by 2,144
  8. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,235
  9. 2018Spider: A Large-Scale Human-Labeled Dataset for Complex and Cross-Domain Semantic Parsing and Text-to-SQL Taskcited by 1,120
  10. 2023A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivitycited by 888
  11. 2023Whose Opinions Do Language Models Reflect?cited by 759
  12. 2022Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learnedcited by 702
10 more
  1. 2023Capabilities of GPT-4 on Medical Challenge Problemscited by 630
  2. 2022News Summarization and Evaluation in the Era of GPT-3cited by 324
  3. 2023How Language Model Hallucinations Can Snowballcited by 310
  4. 2023Evaluating the Logical Reasoning Ability of ChatGPT and GPT-4cited by 202
  5. 2023Can we trust the evaluation on ChatGPT?cited by 63
  6. 2023An Independent Evaluation of ChatGPT on Mathematical Word Problems (MWP)cited by 31
  7. 2022LMentry: A Language Model Benchmark of Elementary Language Taskscited by 21
  8. 2021Did the Model Change? Efficiently Assessing Machine Learning API Shiftscited by 13
  9. 2023ChatLog: Carefully Evaluating the Evolution of ChatGPT Across Timecited by 12
  10. 2022HAPI: A Large-scale Longitudinal Dataset of Commercial ML API Predictionscited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 64; 2024: 78; 2025: 49; 2026: 74.

265 papers.

Sort:

By area:

  1. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,800
  2. 2023GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Promptscited by 520
  3. 2023GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Ciphercited by 401
  4. 2023Low-Resource Languages Jailbreak GPT-4cited by 290
  5. 2023Challenges and Applications of Large Language Modelscited by 251
  6. 2024Generative Representational Instruction Tuningcited by 227
  7. 2024Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMscited by 224
  8. 2023Multilingual Jailbreak Challenges in Large Language Modelscited by 222
  9. 2023Evaluating Large Language Models: A Comprehensive Surveycited by 200
  10. 2023OctoPack: Instruction Tuning Code Large Language Modelscited by 178
  11. 2024Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Toolscited by 170
  12. 2023Art or Artifice? Large Language Models and the False Promise of Creativitycited by 168
38 more
  1. 2023Sociotechnical Safety Evaluation of Generative AI Systemscited by 154
  2. 2023GEMBA-MQM: Detecting Translation Quality Error Spans with GPT-4cited by 143
  3. 2024OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Datasetcited by 143
  4. 2024Who Validates the Validators? Aligning LLM-Assisted Evaluation of LLM Outputs with Human Preferencescited by 142
  5. 2023Generative AIcited by 119
  6. 2023Deception Abilities Emerged in Large Language Modelscited by 99
  7. 2023Fly-Swat or Cannon? Cost-Effective Language Model Choice via Meta-Modelingcited by 94
  8. 2023Where Would I Go Next? Large Language Models as Human Mobility Predictorscited by 92
  9. 2024A Systematic Survey and Critical Review on Evaluating Large Language Models: Challenges, Limitations, and Recommendationscited by 84
  10. 2024Safeguarding Large Language Models: A Surveycited by 73
  11. 2023Split and Merge: Aligning Position Biases in LLM-based Evaluatorscited by 71
  12. 2023Ignore This Title and HackAPrompt: Exposing Systemic Vulnerabilities of LLMs through a Global Scale Prompt Hacking Competitioncited by 65
  13. 2024Building Guardrails for Large Language Modelscited by 61
  14. 2023Self-Guard: Empower the LLM to Safeguard Itselfcited by 52
  15. 2025Demonstrating specification gaming in reasoning modelscited by 48
  16. 2024MEGAnno+: A Human-LLM Collaborative Annotation Systemcited by 47
  17. 2023Theory of Mind in Large Language Models: Examining Performance of 11 State-of-the-Art models vs. Children Aged 7-10 on Advanced Testscited by 46
  18. 2023Large Language Models for Telecom: Forthcoming Impact on the Industrycited by 44
  19. 2024Evaluating Language Model Agency through Negotiationscited by 43
  20. 2024Evolving Code with A Large Language Modelcited by 42
  21. 2023Fighting Fire with Fire: Can ChatGPT Detect AI-generated Text?cited by 39
  22. 2024Cascade Reward Sampling for Efficient Decoding-Time Alignmentcited by 38
  23. 2023Thought Propagation: An Analogical Approach to Complex Reasoning with Large Language Modelscited by 37
  24. 2024Prompt Optimization with EASE? Efficient Ordering-aware Automated Selection of Exemplarscited by 37
  25. 2024Automatic Pseudo-Harmful Prompt Generation for Evaluating False Refusals in Large Language Modelscited by 35
  26. 2025Are You Getting What You Pay For? Auditing Model Substitution in LLM APIscited by 35
  27. 2023Open, Closed, or Small Language Models for Text Classification?cited by 34
  28. 2024Best Practices for Text Annotation with Large Language Modelscited by 34
  29. 2024SOEN-101: Code Generation by Emulating Software Process Models Using Large Language Model Agentscited by 34
  30. 2023Skills-in-Context Prompting: Unlocking Compositionality in Large Language Modelscited by 33
  31. 2023Is GPT-4 a reliable rater? Evaluating Consistency in GPT-4 Text Ratingscited by 33
  32. 2025What Prompts Don't Say: Understanding and Managing Underspecification in LLM Promptscited by 32
  33. 2023Biomedical knowledge graph-optimized prompt generation for large language modelscited by 30
  34. 2024Towards AI-Native Software Engineering (SE 3.0): A Vision and a Challenge Roadmapcited by 30
  35. 2025Generalization Bias in Large Language Model Summarization of Scientific Researchcited by 30
  36. 2023CHATREPORT: Democratizing Sustainability Disclosure Analysis through LLM-based Toolscited by 29
  37. 2024Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Biascited by 28
  38. 2023Experimental Narratives: A Comparison of Human Crowdsourced Storytelling and AI Storytellingcited by 26

50 most cited of 265 All 265 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023GPTEval: A Survey on Assessments of ChatGPT and GPT-4
  2. 2023Summary of ChatGPT-Related Research and Perspective Towards the Future of Large Language Models
  3. 2023GPT-4 Can't Reason
  4. 2023Reliability Check: An Analysis of GPT-3's Response to Sensitive Topics and Prompt Wording
  5. 2024ChatGPT Alternative Solutions: Large Language Models Survey
  6. 2022Prompting GPT-3 To Be Reliable
  7. 2023GPT-3.5, GPT-4, or BARD? Evaluating LLMs Reasoning Ability in Zero-Shot Setting and Performance Boosting Through Prompts
  8. 2024Testing and Evaluation of Large Language Models: Correctness, Non-Toxicity, and Fairness