mlsanityData to 14 Oct 2026

Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiency

arXiv 2506.08343 ↗First posted 10 Jun 2025cs.CLcited by 66

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

Abstract (arXiv)

Recent advances in large reasoning models have enabled complex, step-by-step reasoning but often introduce significant overthinking, resulting in verbose and redundant outputs that hinder efficiency. In this study, we examine whether explicit self-reflection, signaled by tokens such as "Wait" and "Hmm", is necessary for advanced reasoning. We propose NoWait, a simple yet effective approach that disables explicit self-reflection by suppressing these tokens during inference. Extensive experiments on ten benchmarks across textual, visual, and video reasoning tasks show that NoWait reduces chain-of-thought trajectory length by up to 27%-51% in five R1-style model series, without compromising model utility. NoWait thus offers a plug-and-play solution for efficient and utility-preserving multimodal reasoning.

Key facts. Cited by 66 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 30 of its 48 references that are in the map. First posted on arXiv 10 Jun 2025.

Builds on

30 of 48 references.

Closest ancestors · 14

Same direction, newest first

  1. 2025Dynamic Early Exit in Reasoning Models22 Apr
  2. 2025Sleep-time Compute: Beyond Inference Scaling at Test-time17 Apr
  3. 2025Reasoning Models Can Be Effective Without Thinking14 Apr
  4. 2025Speculative Thinking: Enhancing Small-Model Reasoning with Large Model Guidance at Inference Time12 Apr
  5. 2025Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification7 Apr
  6. 2025Understanding Aha Moments: from External Observations to Internal Mechanisms3 Apr
  7. 2025DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models6 Mar
  8. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuning13 Feb
  9. 2025The Danger of Overthinking: Examining the Reasoning-Action Dilemma in Agentic Tasks12 Feb
  10. 2025s1: Simple test-time scaling31 Jan
  11. 2025Reward-Guided Speculative Decoding for Efficient LLM Reasoning31 Jan
  12. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
2 more
  1. 2024Compressed Chain of Thought: Efficient Reasoning Through Dense Representations17 Dec
  2. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug

Other references · 16

  1. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  2. 2023MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGIcited by 2,080
  3. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contextscited by 1,614
  4. 2025Understanding R1-Zero-Like Training: A Critical Perspectivecited by 1,345
  5. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 991
  6. 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?cited by 982
  7. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Modelscited by 665
  8. 2024Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spacescited by 611
  9. 2024MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmarkcited by 512
  10. 2025Video-R1: Reinforcing Video Reasoning in MLLMscited by 484
  11. 2025R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimizationcited by 301
  12. 2025Chain of Draft: Thinking Faster by Writing Lesscited by 297
4 more
  1. 2024Distilling System 2 into System 1cited by 159
  2. 2025Can MLLMs Reason in Multimodality? EMMA: An Enhanced MultiModal ReAsoning Benchmarkcited by 134
  3. 2024Fast Best-of-N Decoding via Speculative Rejectioncited by 112
  4. 2024Group Robust Preference Optimization in Reward-free RLHFcited by 95

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 32; 2026: 34.

66 papers.

Sort:

By area:

  1. 2026CRISP: Compressed Reasoning via Iterative Self-Policy Distillationcited by 58
  2. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Modelscited by 37
  3. 2025OptimalThinkingBench: Evaluating Over and Underthinking in LLMscited by 35
  4. 2025RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMscited by 34
  5. 2025What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoTcited by 25
  6. 2026LLM Reasoning as Trajectories: Step-Specific Representation Geometry and Correctness Signalscited by 24
  7. 2025Trace Length is a Simple Uncertainty Signal in Reasoning Modelscited by 19
  8. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  9. 2025Skywork-R1V3 Technical Reportcited by 16
  10. 2025DRPO: Efficient Reasoning via Decoupled Reward Policy Optimizationcited by 16
  11. 2025Reconsidering Overthinking: Penalizing Internal and External Redundancy in CoT Reasoningcited by 13
  12. 2025Test-time Prompt Interventioncited by 12
38 more
  1. 2025From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Modelscited by 12
  2. 2026Quantifying the Gap between Understanding and Generation within Unified Multimodal Modelscited by 12
  3. 2025ConciseHint: Boosting Efficient Reasoning via Continuous Concise Hints during Generationcited by 10
  4. 2026The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?cited by 10
  5. 2025RAVR: Reference-Answer-guided Variational Reasoning for Large Language Modelscited by 8
  6. 2026Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Modelscited by 8
  7. 2025Are Large Reasoning Models Interruptible?cited by 7
  8. 2026ORBIT: On-policy Exploration-Exploitation for Controllable Multi-Budget Reasoningcited by 7
  9. 2025Meta-R1: Empowering Large Reasoning Models with Metacognitioncited by 6
  10. 2025Think in Parallel, Answer as One: Logit Averaging for Open-Ended Reasoningcited by 6
  11. 2025LYNX: Learning Dynamic Exits for Confidence-Controlled Reasoningcited by 6
  12. 2026Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggerscited by 6
  13. 2026Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compressioncited by 6
  14. 2026Mitigating Overthinking in Large Reasoning Language Models via Reasoning Path Deviation Monitoringcited by 6
  15. 2026When Is Thinking Enough? Early Exit via Sufficiency Assessment for Efficient Reasoningcited by 6
  16. 2025StyleBench: Evaluating thinking styles in Large Language Modelscited by 5
  17. 2025PEAR: Phase Entropy Aware Reward for Efficient Reasoningcited by 5
  18. 2025From "Aha Moments" to Controllable Thinking: Toward Meta-Cognitive Reasoning in Large Reasoning Models via Decoupled Reasoning and Controlcited by 4
  19. 2025Stop Before You Fail: Operational Capability Boundaries for Mitigating Unproductive Reasoning in Large Reasoning Modelscited by 4
  20. 2025Overthinking Reduction with Decoupled Rewards and Curriculum Data Schedulingcited by 4
  21. 2025Explore Briefly, Then Decide: Mitigating LLM Overthinking via Cumulative Entropy Regulationcited by 4
  22. 2026Steering Large Reasoning Models towards Concise Reasoning via Flow Matchingcited by 4
  23. 2026Efficient Reasoning with Balanced Thinkingcited by 4
  24. 2026Thinking with Reasoning Skills: Fewer Tokens, More Accuracycited by 4
  25. 2025MixReasoning: Switching Modes to Thinkcited by 3
  26. 2025ThinkPilot: Steering Reasoning Models via Automated Think-prefixes Optimizationcited by 3
  27. 2025ReflCtrl: Controlling LLM Reflection Efficiently via Representation Engineeringcited by 3
  28. 2026NEAT: Neuron-Based Early Exit for Large Reasoning Modelscited by 3
  29. 2026Learning from Partial Chain-of-Thought via Truncated-Reasoning Self-Distillationcited by 3
  30. 2026Beyond Penalizing Mistakes: Stabilizing Efficiency Training in Large Reasoning Models via Adaptive Correct-Only Rewardscited by 3
  31. 2026Masked Distillation: Internalizing the Chain-of-Thought in Language Modelscited by 3
  32. 2026Recursive Self-Improvement via On-Policy Distillation for Reasoningcited by 3
  33. 2025Think Socially via Cognitive Reasoningcited by 2
  34. 2025Video Reasoning without Trainingcited by 2
  35. 2025Downscaling Intelligence: Exploring Perception and Reasoning Bottlenecks in Small Multimodal Modelscited by 2
  36. 2026From Latent Signals to Reflection Behavior: Tracing Meta-Cognitive Activation Trajectory in R1-Style LLMscited by 2
  37. 2026SAT: Balancing Reasoning Accuracy and Efficiency with Stepwise Adaptive Thinkingcited by 2
  38. 2026Fewer Tokens, Smaller Cache: Reward-Coordinated Efficient Reasoningcited by 2

50 most cited of 66 All 66 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Efficient Reasoning Through Suppression of Self-Affirmation Reflections in Large Reasoning Models
  2. 2025Internal states before wait modulate reasoning patterns
  3. 2026Thinking Past the Answer: Evaluating Harmful Overthinking in Large Reasoning Models
  4. 2026State of Thought Enables Endogenous Reasoning
  5. 2026LatentHarness: Learning Latent Actions for Memory and Reasoning via Counterfactual Policy Distillation
  6. 2025First Try Matters: Revisiting the Role of Reflection in Reasoning Models
  7. 2024Mirror: A Multiple-perspective Self-Reflection Method for Knowledge-rich Reasoning
  8. 2025Double-Checker: Enhancing Reasoning of Slow-Thinking LLMs via Self-Critical Fine-Tuning