mlsanityData to 14 Oct 2026

(Why) Is My Prompt Getting Worse? Rethinking Regression Testing for Evolving LLM APIs

arXiv 2311.11123 ↗First posted 18 Nov 2023cs.SE cs.CLcited by 23

On the map: LLM reasoning → Reasoning for agents, planning and code

Abstract (arXiv)

Large Language Models (LLMs) are increasingly integrated into software applications. Downstream application developers often access LLMs through APIs provided as a service. However, LLM APIs are often updated silently and scheduled to be deprecated, forcing users to continuously adapt to evolving models. This can cause performance regression and affect prompt design choices, as evidenced by our case study on toxicity detection. Based on our case study, we emphasize the need for and re-examine the concept of regression testing for evolving LLM APIs. We argue that regression testing LLMs requires fundamental changes to traditional testing approaches, due to different correctness notions, prompting brittleness, and non-determinism in LLM APIs.

Key facts. Cited by 23 papers in the map (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 19 of its 43 references that are in the map. First posted on arXiv 18 Nov 2023.

Builds on

19 of 43 references.

References · 19

  1. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,234
  2. 2021Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivitycited by 1,135
  3. 2020Beyond Accuracy: Behavioral Testing of NLP models with CheckListcited by 1,006
  4. 2022InCoder: A Generative Model for Code Infilling and Synthesiscited by 498
  5. 2023How is ChatGPT's behavior changing over time?cited by 270
  6. 2022Towards a Unified Multi-Dimensional Evaluator for Text Generationcited by 264
  7. 2018Stress Test Evaluation for Natural Language Inferencecited by 213
  8. 2022Domino: Discovering Systematic Errors with Cross-Modal Embeddingscited by 156
  9. 2020Debugging Tests for Model Explanationscited by 147
  10. 2023An Empirical Study of the Non-determinism of ChatGPT in Code Generationcited by 137
  11. 2017Deceiving Google's Perspective API Built for Detecting Toxic Commentscited by 102
  12. 2019Automatic Testing and Improvement of Machine Translationcited by 36
7 more
  1. 2022Toxicity Detection with Generative Prompt-based Inferencecited by 35
  2. 2023Zeno: An Interactive Framework for Behavioral Evaluation of Machine Learningcited by 33
  3. 2019Mitigating Uncertainty in Document Classificationcited by 22
  4. 2023PromptAid: Prompt Exploration, Perturbation, Testing and Iteration using Visual Analytics for Large Language Modelscited by 19
  5. 2023Beyond Testers' Biases: Guiding Model Testing with Knowledge Bases using LLMscited by 9
  6. 2019Losing Confidence in Quality: Unspoken Evolution of Computer Vision Servicescited by 3
  7. 2020Beware the evolving 'intelligent' web service! An integration architecture tactic to guard AI-first componentscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 6; 2025: 8; 2026: 9.

23 papers.

Sort:

By area:

  1. 2025SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Modelscited by 34
  2. 2025What Prompts Don't Say: Understanding and Managing Underspecification in LLM Promptscited by 32
  3. 2024Prompts Are Programs Too! Understanding How Developers Build Software Containing Promptscited by 19
  4. 2024MUSCLE: A Model Update Strategy for Compatible LLM Evolutioncited by 15
  5. 2025Reflections on the Reproducibility of Commercial LLM Performance in Empirical Software Engineering Studiescited by 12
  6. 2024Beyond the Comfort Zone: Emerging Solutions to Overcome Challenges in Integrating LLMs into Software Productscited by 11
  7. 2025PromptPex: Automatic Test Generation for Language Model Promptscited by 9
  8. 2026ParetoBandit: Budget-Paced Adaptive Routing for Non-Stationary LLM Servingcited by 4
  9. 2024DefectTwin: When LLM Meets Digital Twin for Railway Defect Inspectioncited by 3
  10. 2024From Cool Demos to Production-Ready FMware: Core Challenges and a Technology Roadmapcited by 2
  11. 2025You've Changed: Detecting Modification of Black-Box Large Language Modelscited by 2
  12. 2026Test Before You Deploy: Governing Updates in the LLM Supply Chaincited by 1
11 more
  1. 2024What Is Wrong with My Model? Identifying Systematic Problems with Semantic Data Slicingcited by 0
  2. 2025The Hitchhikers Guide to Production-ready Trustworthy Foundation Model powered Software (FMware)cited by 0
  3. 2025Delving Into the Psychology of Machines: Exploring the Structure of Self-Regulated Learning via LLM-Generated Survey Responsescited by 0
  4. 2025ReCatcher: Towards LLMs Regression Testing for Code Generationcited by 0
  5. 2026Results-Actionability Gap: Understanding How Practitioners Evaluate LLM Products in the Wildcited by 0
  6. 2026What Aggregate Scores Miss: Measuring Item-Level Regressions in Commercial LLM API Migrationscited by 0
  7. 2026UpgradeBench: A Decision-Centric Benchmark for Upgrading Fine-Tuned LLM Specialistscited by 0
  8. 2026A collective capability boundary in frontier large language models on guideline-conformant and case-specific oncology decision-makingcited by 0
  9. 2026How Reproducible Are Evaluation Conclusions? A Self-Audit of LLM-Inferred Prompt Structurecited by 0
  10. 2026Large Language Models, Machine-Assisted Cultural Analytics, and the Agentic Turncited by 0
  11. 2026Beyond Aggregate Accuracy: Item-Level Transitions in Prompt Optimizationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Challenges in Testing Large Language Model Based Software: A Faceted Taxonomy
  2. 2025A Taxonomy of Prompt Defects in LLM Systems
  3. 2026When Generic Prompt Improvements Hurt: Evaluation-Driven Iteration for LLM Applications
  4. 2026SWE-Mutation: Can LLMs Generate Reliable Test Suites in Software Engineering?
  5. 2026When LLMs Lag Behind: Knowledge Conflicts from Evolving APIs in Code Generation
  6. 2025Rethinking Testing for LLM Applications: Characteristics, Challenges, and a Lightweight Interaction Protocol
  7. 2026LLMs taking shortcuts in test generation: A study with SAP HANA and LevelDB
  8. 2025Artificial or Just Artful? Do LLMs Bend the Rules in Programming? ↗