mlsanityData to 7 Oct 2026

From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step

arXiv 2405.14838 ↗First posted 23 May 2024cs.CL cs.AI cs.LGcited by 225

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

When leveraging language models for reasoning tasks, generating explicit chain-of-thought (CoT) steps often proves essential for achieving high accuracy in final outputs. In this paper, we investigate if models can be taught to internalize these CoT steps. To this end, we propose a simple yet effective method for internalizing CoT steps: starting with a model trained for explicit CoT reasoning, we gradually remove the intermediate steps and finetune the model. This process allows the model to internalize the intermediate reasoning steps, thus simplifying the reasoning process while maintaining high performance. Our approach enables a GPT-2 Small model to solve 9-by-9 multiplication with up to 99% accuracy, whereas standard training cannot solve beyond 4-by-4 multiplication. Furthermore, our method proves effective on larger language models, such as Mistral 7B, achieving over 50% accuracy on GSM8K without producing any intermediate steps.

Key facts. Cited by 225 papers in the map (data through 7 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 10 of its 16 references that are in the map. First posted on arXiv 23 May 2024.

Builds on

10 of 16 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Beyond A*: Better Planning with Transformers via Search Dynamics Bootstrapping21 Feb
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Models28 Jan

Other references · 8

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,970
  4. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,883
  5. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Modelscited by 871
  6. 2019Designing and Interpreting Probes with Control Taskscited by 608
  7. 2023Faith and Fate: Limits of Transformers on Compositionalitycited by 496
  8. 2023GPT Can Solve Mathematical Problems Without a Calculatorcited by 68

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 21; 2025: 100; 2026: 104.

225 papers.

Sort:

By area:

  1. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 763
  2. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Modelscited by 456
  3. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 381
  4. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 372
  5. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscationcited by 338
  6. 2024To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoningcited by 257
  7. 2025TokenSkip: Controllable Chain-of-Thought Compression in LLMscited by 246
  8. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 216
  9. 2024Compressed Chain of Thought: Efficient Reasoning Through Dense Representationscited by 192
  10. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuningcited by 167
  11. 2024Distilling System 2 into System 1cited by 152
  12. 2024LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMscited by 142
38 more
  1. 2025A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyondcited by 131
  2. 2025Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokenscited by 120
  3. 2024On Memorization of Large Language Models in Logical Reasoningcited by 114
  4. 2025A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systemscited by 108
  5. 2025LightThinker: Thinking Step-by-Step Compressioncited by 104
  6. 2025Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Modelscited by 94
  7. 2025Efficient Reasoning Models: A Surveycited by 93
  8. 2025Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoningcited by 87
  9. 2025Think Silently, Think Fast: Dynamic Latent Compression of LLM Reasoning Chainscited by 81
  10. 2025Efficient Reasoning with Hidden Thinkingcited by 72
  11. 2025Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoningcited by 72
  12. 2025A Survey on Latent Reasoningcited by 71
  13. 2025Are DeepSeek R1 And Other Reasoning Models More Faithful?cited by 69
  14. 2025SIM-CoT: Supervised Implicit Chain-of-Thoughtcited by 63
  15. 2024Can Language Models Learn to Skip Steps?cited by 59
  16. 2025Efficient Inference for Large Reasoning Models: A Surveycited by 58
  17. 2025Think with 3D: Geometric Imagination Grounded Spatial Reasoning from Limited Viewscited by 58
  18. 2025Measuring Chain of Thought Faithfulness by Unlearning Reasoning Stepscited by 43
  19. 2025Reasoning to Learn from Latent Thoughtscited by 43
  20. 2025Implicit Reasoning in Large Language Models: A Comprehensive Surveycited by 42
  21. 2024Optima: Optimizing Effectiveness and Efficiency for LLM-Based Multi-Agent Systemcited by 40
  22. 2024Dualformer: Controllable Fast and Slow Thinking by Learning with Randomized Reasoning Tracescited by 40
  23. 2025Seek in the Dark: Reasoning via Test-Time Instance-Level Policy Gradient in Latent Spacecited by 40
  24. 2025Making Slow Thinking Faster: Compressing LLM Chain-of-Thought via Step Entropycited by 37
  25. 2025LLM Latent Reasoning as Chain of Superpositioncited by 37
  26. 2025Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Modelscited by 35
  27. 2025THOUGHTTERMINATOR: Benchmarking, Calibrating, and Mitigating Overthinking in Reasoning Modelscited by 34
  28. 2025LLMs are Single-threaded Reasoners: Demystifying the Working Mechanism of Soft Thinkingcited by 34
  29. 2024Break the Chain: Large Language Models Can be Shortcut Reasonerscited by 32
  30. 2025Continuous Chain of Thought Enables Parallel Exploration and Reasoningcited by 32
  31. 2025Self-Improving Transformers Overcome Easy-to-Hard and Length Generalization Challengescited by 30
  32. 2025Hybrid Latent Reasoning via Reinforcement Learningcited by 30
  33. 2025Lower Bounds for Chain-of-Thought Reasoning in Hard-Attention Transformerscited by 28
  34. 2026LaST0: Latent Spatio-Temporal Chain-of-Thought for Robotic Vision-Language-Action Modelcited by 28
  35. 2025SPEED-RL: Faster Training of Reasoning Models via Online Curriculum Learningcited by 27
  36. 2024Deliberation in Latent Space via Differentiable Cache Augmentationcited by 25
  37. 2025VeriThinker: Learning to Verify Makes Reasoning Model Efficientcited by 24
  38. 2025Thought Manipulation: External Thought Can Be Efficient for Large Reasoning Modelscited by 23

50 most cited of 225 All 225 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning
  2. 2025The Challenge of Teaching Reasoning to LLMs Without RL or Distillation
  3. 2025Efficient Long CoT Reasoning in Small Language Models
  4. 2022Large Language Models Are Reasoning Teachers
  5. 2025SCOUT: Teaching Pre-trained Language Models to Enhance Reasoning via Flow Chain-of-Thought
  6. 2022Distilling Reasoning Capabilities into Smaller Language Models
  7. 2025Innate Reasoning is Not Enough: In-Context Learning Enhances Reasoning Large Language Models with Less Overthinking
  8. 2025Unveiling the Key Factors for Distilling Chain-of-Thought Reasoning