mlsanityData to 14 Oct 2026

SynAdapt: Learning Adaptive Reasoning in Large Language Models via Synthetic Continuous Chain-of-Thought

arXiv 2508.00574 ↗First posted 1 Aug 2025cs.CL cs.AIcited by 13

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

Abstract (arXiv)

While Chain-of-Thought (CoT) reasoning improves model performance, it incurs significant time costs due to the generation of discrete CoT tokens (DCoT). Continuous CoT (CCoT) offers a more efficient alternative, but existing CCoT methods are hampered by indirect fine-tuning, limited alignment, or inconsistent targets. To overcome these limitations, we propose textitSynAdapt, an innovative efficient reasoning framework. Specifically, textitSynAdapt generates the synthetic CCoT to serve as a precise and effective alignment target for LLMs. This synthetic CCoT explicitly guides the LLM to learn CCoT and derive accurate answers directly. Furthermore, relying solely on CCoT is insufficient for solving hard questions. To address this, textitSynAdapt integrates a difficulty classifier that leverages both question context and CCoT to identify hard questions. CCoT can effectively help identify hard questions after some brief reasoning. We then adaptively prompt the LLM to re-think these hard questions for improved performance. Extensive experimental results across various benchmarks from different difficulty levels strongly demonstrate the effectiveness of our method, achieving the best accuracy-efficiency trade-off.

Key facts. Cited by 13 papers in the map (data through 14 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 42 of its 54 references that are in the map. First posted on arXiv 1 Aug 2025.

Builds on

42 of 54 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025DART: Distilling Autoregressive Reasoning to Silent Thought13 Jun
  2. 2025Reasoning by Superposition: A Theoretical Perspective on Chain of Continuous Thought18 May
  3. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation28 Feb
  4. 2025Chain of Draft: Thinking Faster by Writing Less25 Feb
  5. 2025SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs17 Feb
  6. 2025Enhancing Auto-regressive Chain-of-Thought through Loop-Aligned Reasoning12 Feb
  7. 2024Training Large Language Models to Reason in a Continuous Latent Space9 Dec
  8. 2024Distilling System 2 into System 18 Jul
  9. 2024From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step23 May
  10. 2024Let's Think Dot by Dot: Hidden Computation in Transformer Language Models24 Apr
  11. 2024Chain of Thought Empowers Transformers to Solve Inherently Serial Problems20 Feb
  12. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Models21 May

Other references · 30

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2023Let's Verify Step by Stepcited by 3,789
  5. 2024Mixtral of Expertscited by 1,823
  6. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,071
  7. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 682
  8. 2023The Internal State of an LLM Knows When It's Lyingcited by 659
  9. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learningcited by 384
  10. 2024LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compressioncited by 348
  11. 2017Six Challenges for Neural Machine Translationcited by 335
  12. 2025Demystifying Long Chain-of-Thought Reasoning in LLMscited by 326
18 more
  1. 2025DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoningcited by 280
  2. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruningcited by 279
  3. 2023Think before you speak: Training Language Models With Pause Tokenscited by 264
  4. 2023Challenges and Applications of Large Language Modelscited by 254
  5. 2025TokenSkip: Controllable Chain-of-Thought Compression in LLMscited by 249
  6. 2024Token-Budget-Aware LLM Reasoningcited by 247
  7. 2025Training Language Models to Reason Efficientlycited by 247
  8. 2025Reasoning with Latent Thoughts: On the Power of Looped Transformerscited by 240
  9. 2024Compressed Chain of Thought: Efficient Reasoning Through Dense Representationscited by 201
  10. 2025Reasoning Models Can Be Effective Without Thinkingcited by 187
  11. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuningcited by 169
  12. 2025DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Modelscited by 158
  13. 2024Generalization or Memorization: Data Contamination and Trustworthy Evaluation for Large Language Modelscited by 122
  14. 2025Self-Training Elicits Concise Reasoning in Large Language Modelscited by 99
  15. 2025How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approachcited by 97
  16. 2024Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM's Reasoning Capabilitycited by 83
  17. 2024Factual Confidence of LLMs: on Reliability and Robustness of Current Estimatorscited by 61
  18. 2024A Survey on Medical Large Language Models: Technology, Application, Trustworthiness, and Future Directionscited by 34

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 10.

13 papers.

Sort:

By area:

  1. 2025Monet: Reasoning in Latent Visual Space Beyond Images and Languagecited by 82
  2. 2025Implicit Reasoning in Large Language Models: A Comprehensive Surveycited by 43
  3. 2025SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimizationcited by 21
  4. 2026LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoningcited by 8
  5. 2026From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AIcited by 6
  6. 2026Capabilities and Fundamental Limits of Latent Chain-of-Thoughtcited by 4
  7. 2026Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Modelcited by 3
  8. 2026ImgCoT: Compressing Long Chain of Thought into Compact Visual Tokens for Efficient Reasoning of Large Language Modelcited by 1
  9. 2026Multimodal Continuous Reasoning via Asymmetric Mutual Variational Learningcited by 1
  10. 2026LatentRAG: Latent Reasoning and Retrieval for Efficient Agentic RAGcited by 0
  11. 2026RuPLaR : Efficient Latent Compression of LLM Reasoning Chains with Rule-Based Priors From Multi-Step to One-Stepcited by 0
  12. 2026N-GRPO: Embedding-Level Neighbor Mixing for Enhanced Policy Optimizationcited by 0
1 more
  1. 2026Think Wider: Mitigating Latent Rank Collapse in Implicit Chain-of-Thought Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Early Stopping Chain-of-thoughts in Large Language Models
  2. 2025Rethinking the Generation of High-Quality CoT Data from the Perspective of LLM-Adaptive Question Difficulty Grading
  3. 2025CoT-X: An Adaptive Framework for Cross-Model Chain-of-Thought Transfer and Optimization
  4. 2026CoT-Core: Accelerating LLM Evaluation via CoT-Aware Coreset Selection
  5. 2024On the Hardness of Faithful Chain-of-Thought Reasoning in Large Language Models
  6. 2024Fine-Tuning on Diverse Reasoning Chains Drives Within-Inference CoT Refinement in LLMs
  7. 2026Tandem: Riding Together with Large and Small Language Models for Efficient Reasoning
  8. 2023ThoughtSource: A central hub for large language model reasoning data