mlsanityData to 7 Oct 2026

AIR: Complex Instruction Generation via Automatic Iterative Refinement

arXiv 2502.17787 ↗First posted 25 Feb 2025cs.CL cs.AIcited by 10

Not on the map. Why?

Abstract (arXiv)

With the development of large language models, their ability to follow simple instructions has significantly improved. However, adhering to complex instructions remains a major challenge. Current approaches to generating complex instructions are often irrelevant to the current instruction requirements or suffer from limited scalability and diversity. Moreover, methods such as back-translation, while effective for simple instruction generation, fail to leverage the rich contents and structures in large web corpora. In this paper, we propose a novel automatic iterative refinement framework to generate complex instructions with constraints, which not only better reflects the requirements of real scenarios but also significantly enhances LLMs' ability to follow complex instructions. The AIR framework consists of two stages: (1)Generate an initial instruction from a document; (2)Iteratively refine instructions with LLM-as-judge guidance by comparing the model's output with the document to incorporate valuable constraints. Finally, we construct the AIR-10K dataset with 10K complex instructions and demonstrate that instructions generated with our approach significantly improve the model's ability to follow complex instructions, outperforming existing methods for instruction generation.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 30 of its 46 references that are in the map. First posted on arXiv 25 Feb 2025.

Builds on

30 of 46 references.

References · 30

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  5. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  6. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
  7. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  8. 2018CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledgecited by 1,793
  9. 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,442
  10. 2023LIMA: Less Is More for Alignmentcited by 1,080
  11. 2024Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluatorscited by 865
  12. 2023Enhancing Chat Language Models by Scaling High-quality Instructional Conversationscited by 825
18 more
  1. 2015Improving Neural Machine Translation Models with Monolingual Datacited by 784
  2. 2024Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Researchcited by 475
  3. 2023How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resourcescited by 405
  4. 2023What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuningcited by 354
  5. 2024Justice or Prejudice? Quantifying Biases in LLM-as-a-Judgecited by 354
  6. 2024Benchmarking Complex Instruction-Following with Multiple Constraints Compositioncited by 147
  7. 2023ISR-LLM: Iterative Self-Refined Large Language Model for Long-Horizon Sequential Task Planningcited by 77
  8. 2023Can Large Language Models Understand Real-World Complex Instructions?cited by 69
  9. 2024CFBench: A Comprehensive Constraints-Following Benchmark for LLMscited by 65
  10. 2024Iterative Refinement of Project-Level Code Context for Precise Code Generation with Compiler Feedbackcited by 45
  11. 2024Suri: Multi-constraint Instruction Following for Long-form Text Generationcited by 34
  12. 2023MUFFIN: Curating Multi-Faceted Instructions for Improving Instruction-Followingcited by 30
  13. 2024I-SHEEP: Self-Alignment of LLM from Scratch through an Iterative Self-Enhancement Paradigmcited by 27
  14. 2024Human-Instruction-Free LLM Self-Alignment with Limited Samplescited by 25
  15. 2024Better Alignment with Instruction Back-and-Forth Translationcited by 16
  16. 2025MuSC: Improving Complex Instruction Following with Multi-granularity Self-Contrastive Trainingcited by 12
  17. 2024Kun: Answer Polishment for Chinese Self-Alignment with Instruction Back-Translationcited by 9
  18. 20242D-DPO: Scaling Direct Preference Optimization with 2-Dimensional Supervisioncited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 6; 2026: 4.

10 papers.

  1. 2025Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning?cited by 43
  2. 2025Think-J: Learning to Think for Generative LLM-as-a-Judgecited by 20
  3. 2025Incentivizing Reasoning for Advanced Instruction-Following of Large Language Modelscited by 17
  4. 2025RECAST: Expanding the Boundaries of LLMs' Complex Instruction Following with Multi-Constraint Datacited by 11
  5. 2025Replay Failures as Successes: Sample-Efficient Reinforcement Learning for Instruction Followingcited by 7
  6. 2026ELLMob: Event-Driven Human Mobility Generation with Self-Aligned LLM Frameworkcited by 3
  7. 2025DecIF: Improving Instruction-Following through Meta-Decompositioncited by 1
  8. 2026Enhancing LLM Instruction Following: An Evaluation-Driven Multi-Agentic Workflow for Prompt Instructions Optimizationcited by 1
  9. 2026Large Language Models Can Follow Instructions, But Not Many at Once: Phase Transitions in Compositional Constraint Satisfactioncited by 0
  10. 2026ADE: Agentic Data Evolution Framework for Human-Centered Objectivescited by 0