mlsanityData to 14 Oct 2026

PAL: Program-aided Language Models

arXiv 2211.10435 ↗First posted 18 Nov 2022cs.CL cs.AIcited by 652

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Large language models (LLMs) have recently demonstrated an impressive ability to perform arithmetic and symbolic reasoning tasks, when provided with a few examples at test time ("few-shot prompting"). Much of this success can be attributed to prompting methods such as "chain-of-thought'', which employ LLMs for both understanding the problem description by decomposing it into steps, as well as solving each step of the problem. While LLMs seem to be adept at this sort of step-by-step decomposition, LLMs often make logical and arithmetic mistakes in the solution part, even when the problem is decomposed correctly. In this paper, we present Program-Aided Language models (PAL): a novel approach that uses the LLM to read natural language problems and generate programs as the intermediate reasoning steps, but offloads the solution step to a runtime such as a Python interpreter. With PAL, decomposing the natural language problem into runnable steps remains the only learning task for the LLM, while solving is delegated to the interpreter. We demonstrate this synergy between a neural LLM and a symbolic interpreter across 13 mathematical, symbolic, and algorithmic reasoning tasks from BIG-Bench Hard and other benchmarks. In all these natural language reasoning tasks, generating code using an LLM and reasoning using a Python interpreter leads to more accurate results than much larger models. For example, PAL using Codex achieves state-of-the-art few-shot accuracy on the GSM8K benchmark of math word problems, surpassing PaLM-540B which uses chain-of-thought by absolute 15% top-1. Our code and data are publicly available at http://reasonwithpal.com/ .

Key facts. Cited by 652 papers in the map (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 46 of its 67 references that are in the map. First posted on arXiv 18 Nov 2022.

Builds on

46 of 67 references.

Closest ancestors · 4

Same direction, newest first

  1. 2022Language Models of Code are Few-Shot Commonsense Learners13 Oct
  2. 2022Measuring and Narrowing the Compositionality Gap in Language Models7 Oct
  3. 2022ReAct: Synergizing Reasoning and Acting in Language Models6 Oct
  4. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Models30 Nov

Other references · 42

  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. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  5. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  6. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  7. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,326
  8. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  9. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,780
  10. 2019The Curious Case of Neural Text Degenerationcited by 2,916
  11. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,348
  12. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,234
30 more
  1. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,998
  2. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
  3. 2021Multitask Prompted Training Enables Zero-Shot Task Generalizationcited by 1,470
  4. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Modelscited by 1,407
  5. 2022Solving Quantitative Reasoning Problems with Language Modelscited by 1,392
  6. 2021Are NLP Models really able to Solve Simple Math Word Problems?cited by 878
  7. 2019MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalismscited by 662
  8. 2017Program Induction by Rationale Generation : Learning to Solve and Explain Algebraic Word Problemscited by 650
  9. 2021A Diverse Corpus for Evaluating and Developing English Math Word Problem Solverscited by 318
  10. 2022Autoformalization with Large Language Modelscited by 263
  11. 2020Neurosymbolic AI: The 3rd Wavecited by 255
  12. 2021The GEM Benchmark: Natural Language Generation, its Evaluation and Metricscited by 239
  13. 2022Binding Language Models in Symbolic Languagescited by 228
  14. 2018Deep Learning: A Critical Appraisalcited by 217
  15. 2020The Next Decade in AI: Four Steps Towards Robust Artificial Intelligencecited by 195
  16. 2021Constrained Language Models Yield Few-Shot Semantic Parserscited by 163
  17. 2021Investigating the Limitations of Transformers with Simple Arithmetic Taskscited by 139
  18. 2020Injecting Numerical Reasoning Skills into Language Modelscited by 133
  19. 2022Text and Patterns: For Effective Chain of Thought, It Takes Two to Tangocited by 128
  20. 2022Rationale-Augmented Ensembles in Language Modelscited by 111
  21. 2021A Recipe For Arbitrary Text Style Transfer with Large Language Modelscited by 107
  22. 2022Language Model Cascadescited by 107
  23. 2021Improving Coherence and Consistency in Neural Sequence Models with Dual-System, Neuro-Symbolic Reasoningcited by 85
  24. 2022Limitations of Language Models in Arithmetic and Symbolic Inductioncited by 71
  25. 2019Neural Module Networks for Reasoning over Textcited by 66
  26. 2019Giving BERT a Calculator: Finding Operations and Arguments with Reading Comprehensioncited by 51
  27. 2021Few-Shot Semantic Parsing with Language Models Trained On Codecited by 50
  28. 2022Reasoning Like Program Executorscited by 45
  29. 2016Learning Executable Semantic Parsers for Natural Language Understandingcited by 21
  30. 2019Just Add Functions: A Neural-Symbolic Language Modelcited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 10; 2023: 189; 2024: 131; 2025: 149; 2026: 173.

652 papers.

Sort:

By area:

  1. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  2. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  3. 2023A Survey of Large Language Modelscited by 2,576
  4. 2023WebArena: A Realistic Web Environment for Building Autonomous Agentscited by 2,006
  5. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
  6. 2023AgentBench: Evaluating LLMs as Agentscited by 1,267
  7. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,239
  8. 2023HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Facecited by 1,190
  9. 2024DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligencecited by 1,181
  10. 2022Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Taskscited by 1,076
  11. 2023Teaching Large Language Models to Self-Debugcited by 988
  12. 2023A Survey on Multimodal Large Language Modelscited by 909
38 more
  1. 2023Large Language Models Can Be Easily Distracted by Irrelevant Contextcited by 841
  2. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 832
  3. 2023StarCoder: may the source be with you!cited by 799
  4. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 783
  5. 2023C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Modelscited by 708
  6. 2023Safe RLHF: Safe Reinforcement Learning from Human Feedbackcited by 656
  7. 2023ViperGPT: Visual Inference via Python Execution for Reasoningcited by 603
  8. 2022Towards Reasoning in Large Language Models: A Surveycited by 596
  9. 2023Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Modelscited by 574
  10. 2024Executable Code Actions Elicit Better LLM Agentscited by 564
  11. 2023Identifying the Risks of LM Agents with an LM-Emulated Sandboxcited by 548
  12. 2024BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructionscited by 546
  13. 2024Large Language Models: A Surveycited by 525
  14. 2023MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Actioncited by 517
  15. 2023LLM+P: Empowering Large Language Models with Optimal Planning Proficiencycited by 501
  16. 2023MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuningcited by 487
  17. 2024AgentPoison: Red-teaming LLM Agents via Poisoning Memory or Knowledge Basescited by 429
  18. 2024Automated Design of Agentic Systemscited by 407
  19. 2023Augmented Language Models: a Surveycited by 403
  20. 2023ToolAlpaca: Generalized Tool Learning for Language Models with 3000 Simulated Casescited by 393
  21. 2023Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoningcited by 391
  22. 2023Llemma: An Open Language Model For Mathematicscited by 388
  23. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscationcited by 384
  24. 2023Chameleon: Plug-and-Play Compositional Reasoning with Large Language Modelscited by 356
  25. 2023A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesiscited by 353
  26. 2023Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergycited by 346
  27. 2023StructGPT: A General Framework for Large Language Model to Reason over Structured Datacited by 343
  28. 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 320
  29. 2022Large Language Models are Better Reasoners with Self-Verificationcited by 307
  30. 2023Tool Learning with Foundation Modelscited by 302
  31. 2023MINT: Evaluating LLMs in Multi-turn Interaction with Tools and Language Feedbackcited by 297
  32. 2025Defeating Prompt Injections by Designcited by 297
  33. 2023Faithful Chain-of-Thought Reasoningcited by 288
  34. 2024We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?cited by 286
  35. 2023ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solvingcited by 282
  36. 2023ToolQA: A Dataset for LLM Question Answering with External Toolscited by 274
  37. 2024Chain of Thought Empowers Transformers to Solve Inherently Serial Problemscited by 260
  38. 2023SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Modelscited by 256

50 most cited of 652 All 652 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023Program-Aided Reasoners (better) Know What They Know
  2. 2023Large Language Models Are Not Strong Abstract Reasoners
  3. 2026LLMs as ASP Programmers: Self-Correction Enables Task-Agnostic Nonmonotonic Reasoning
  4. 2025Intermediate Languages Matter: Formal Languages and LLMs affect Neurosymbolic Reasoning
  5. 2023Reasoning in Large Language Models Through Symbolic Math Word Problems
  6. 2024Investigating Symbolic Capabilities of Large Language Models
  7. 2023Leveraging Large Language Models to Generate Answer Set Programs
  8. 2024Assessing the Emergent Symbolic Reasoning Abilities of Llama Large Language Models