mlsanityData to 7 Oct 2026

Compressed Chain of Thought: Efficient Reasoning Through Dense Representations

arXiv 2412.13171 ↗First posted 17 Dec 2024cs.CLcited by 192

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Chain-of-thought (CoT) decoding enables language models to improve reasoning performance at the cost of high generation latency in decoding. Recent proposals have explored variants of contemplation tokens, a term we introduce that refers to special tokens used during inference to allow for extra computation. Prior work has considered fixed-length sequences drawn from a discrete set of embeddings as contemplation tokens. Here we propose Compressed Chain-of-Thought (CCoT), a framework to generate contentful and continuous contemplation tokens of variable sequence length. The generated contemplation tokens are compressed representations of explicit reasoning chains, and our method can be applied to off-the-shelf decoder language models. Through experiments, we illustrate how CCoT enables additional reasoning over dense contentful representations to achieve corresponding improvements in accuracy. Moreover, the reasoning improvements can be adaptively modified on demand by controlling the number of contemplation tokens generated.

Key facts. Cited by 192 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 17 of its 18 references that are in the map. First posted on arXiv 17 Dec 2024.

Builds on

17 of 18 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023Fast Chain-of-Thought: A Glance of Future from Parallel Decoding Leads to Answers Faster14 Nov
  2. 2023Think before you speak: Training Language Models With Pause Tokens3 Oct
  3. 2023Skeleton-of-Thought: Prompting LLMs for Efficient Parallel Generation28 Jul
  4. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Models17 May

Other references · 13

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2022Large Language Models are Zero-Shot Reasonerscited by 4,323
  4. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 763
  5. 2023Measuring Faithfulness in Chain-of-Thought Reasoningcited by 668
  6. 2023LLMLingua: Compressing Prompts for Accelerated Inference of Large Language Modelscited by 279
  7. 2024From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Stepcited by 225
  8. 2023CacheGen: KV Cache Compression and Streaming for Fast Large Language Model Servingcited by 199
  9. 2024Let's Think Dot by Dot: Hidden Computation in Transformer Language Modelscited by 186
  10. 2023Implicit Chain of Thought Reasoning via Knowledge Distillationcited by 165
  11. 2023Chain of Hindsight Aligns Language Models with Feedbackcited by 147
  12. 2023Dodo: Dynamic Contextual Compression for Decoder-only LMscited by 23
1 more
  1. 2024Fine-Tuning on Diverse Reasoning Chains Drives Within-Inference CoT Refinement in LLMscited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 102; 2026: 90.

192 papers.

Sort:

By area:

  1. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Modelscited by 456
  2. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 372
  3. 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 319
  4. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruningcited by 278
  5. 2025TokenSkip: Controllable Chain-of-Thought Compression in LLMscited by 246
  6. 2025Dynamic Early Exit in Reasoning Modelscited by 217
  7. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 216
  8. 2025Reasoning Models Can Be Effective Without Thinkingcited by 182
  9. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuningcited by 167
  10. 2025A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyondcited by 131
  11. 2025Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computationcited by 120
  12. 2025SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMscited by 110
38 more
  1. 2025LightThinker: Thinking Step-by-Step Compressioncited by 104
  2. 2025Latent Visual Reasoningcited by 99
  3. 2025Efficient Reasoning Models: A Surveycited by 93
  4. 2025Think Silently, Think Fast: Dynamic Latent Compression of LLM Reasoning Chainscited by 81
  5. 2025Efficient Reasoning with Hidden Thinkingcited by 72
  6. 2025Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoningcited by 72
  7. 2025A Survey on Latent Reasoningcited by 71
  8. 2025Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiencycited by 65
  9. 2025Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokenscited by 65
  10. 2025SIM-CoT: Supervised Implicit Chain-of-Thoughtcited by 63
  11. 2025Efficient Inference for Large Reasoning Models: A Surveycited by 58
  12. 2025A Survey of LLM-Driven AI Agent Communication: Protocols, Security Risks, and Defense Countermeasurescited by 54
  13. 2025InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Modelscited by 53
  14. 2026LoopFormer: Elastic-Depth Looped Transformers for Latent Reasoning via Shortcut Modulationcited by 48
  15. 2025Implicit Reasoning in Large Language Models: A Comprehensive Surveycited by 42
  16. 2025A Survey of Test-Time Compute: From Intuitive Inference to Deliberate Reasoningcited by 40
  17. 2025Seek in the Dark: Reasoning via Test-Time Instance-Level Policy Gradient in Latent Spacecited by 40
  18. 2025SoftCoT++: Test-Time Scaling with Soft Chain-of-Thought Reasoningcited by 38
  19. 2025ARM: Adaptive Reasoning Modelcited by 38
  20. 2025LLM Latent Reasoning as Chain of Superpositioncited by 37
  21. 2025Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Modelscited by 35
  22. 2025Continuous Chain of Thought Enables Parallel Exploration and Reasoningcited by 32
  23. 2025Pruning the Unsurprising: Efficient LLM Reasoning via First-Token Surprisalcited by 32
  24. 2025Reasoning on a Budget: A Survey of Adaptive and Controllable Test-Time Compute in LLMscited by 31
  25. 2025Enabling Agents to Communicate Entirely in Latent Spacecited by 31
  26. 2025R1-Compress: Long Chain-of-Thought Compression via Chunk Compression and Searchcited by 29
  27. 2025FlashThink: An Early Exit Method For Efficient Reasoningcited by 28
  28. 2025Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasonercited by 28
  29. 2026LaST-VLA: Thinking in Latent Spatio-Temporal Space for Vision-Language-Action in Autonomous Drivingcited by 26
  30. 2025LaDiR: Latent Diffusion Enhances LLMs for Text Reasoningcited by 25
  31. 2025Sketch-in-Latents: Eliciting Unified Reasoning in MLLMscited by 24
  32. 2025Thought Manipulation: External Thought Can Be Efficient for Large Reasoning Modelscited by 23
  33. 2025System-1.5 Reasoning: Traversal in Language and Latent Spaces with Dynamic Shortcutscited by 23
  34. 2025Encode, Think, Decode: Scaling test-time reasoning with recursive latent thoughtscited by 23
  35. 2025Activation Steering for Chain-of-Thought Compressioncited by 22
  36. 2025Parallel Continuous Chain-of-Thought with Jacobi Iterationcited by 21
  37. 2025ThinKV: Thought-Adaptive KV Cache Compression for Efficient Reasoning Modelscited by 21
  38. 2025SwiReasoning: Switch-Thinking in Latent and Explicit for Pareto-Superior Reasoning LLMscited by 20

50 most cited of 192 All 192 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024Expediting and Elevating Large Language Model Reasoning via Hidden Chain-of-Thought Decoding
  2. 2026Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compression
  3. 2024C3oT: Generating Shorter Chain-of-Thought without Compromising Effectiveness
  4. 2026HybridThinker: Efficient Chain-of-Thought Reasoning via Compressed Memory and Transient Thought Steps
  5. 2026SYNLAT: Syntax-Aligned Text-Latent Compression for Chain-of-Thought Reasoning
  6. 2026Memory Augmentation Unlocks Efficient Chain-of-Thought Reasoning
  7. 2026Zipping the Thought: When and How Compressed Reasoning Data Works in LLM Post-Training
  8. 2026Demystifying Entropy-based Selection for Chain-of-Thought Compression in Large Reasoning Models