# When to Continue Thinking: Adaptive Thinking Mode Switching for Efficient Reasoning

arXiv: https://arxiv.org/abs/2505.15400 · HTML: https://mlsanity.com/paper/2505.15400/

## Key facts

- Cited by 26 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 28 of its 33 references that are in the map.
- First posted on arXiv 21 May 2025.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 28 of 28)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [s1: Simple test-time scaling](https://mlsanity.com/paper/2501.19393/) (2025)
- [OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems](https://mlsanity.com/paper/2402.14008/) (2024)
- [Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?](https://mlsanity.com/paper/2504.13837/) (2025)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [Training Language Models to Reason Efficiently](https://mlsanity.com/paper/2502.04463/) (2025)
- [Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment](https://mlsanity.com/paper/2308.09662/) (2023)
- [Dynamic Early Exit in Reasoning Models](https://mlsanity.com/paper/2504.15895/) (2025)
- [Self-Evaluation Guided Beam Search for Reasoning](https://mlsanity.com/paper/2305.00633/) (2023)
- [Reasoning Models Can Be Effective Without Thinking](https://mlsanity.com/paper/2504.09858/) (2025)
- [Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification](https://mlsanity.com/paper/2504.05419/) (2025)
- [DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models](https://mlsanity.com/paper/2503.04472/) (2025)
- [Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuning](https://mlsanity.com/paper/2503.07572/) (2025)
- [Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable](https://mlsanity.com/paper/2503.00555/) (2025)
- [C3oT: Generating Shorter Chain-of-Thought without Compromising Effectiveness](https://mlsanity.com/paper/2412.11664/) (2024)
- [Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching](https://mlsanity.com/paper/2503.05179/) (2025)
- [SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs](https://mlsanity.com/paper/2502.12134/) (2025)
- [Reward-Guided Speculative Decoding for Efficient LLM Reasoning](https://mlsanity.com/paper/2501.19324/) (2025)
- [OverThink: Slowdown Attacks on Reasoning LLMs](https://mlsanity.com/paper/2502.02542/) (2025)
- [Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?](https://mlsanity.com/paper/2502.12215/) (2025)
- [Learning to Route LLMs with Confidence Tokens](https://mlsanity.com/paper/2410.13284/) (2024)
- [Unlocking Efficient Long-to-Short LLM Reasoning with Model Merging](https://mlsanity.com/paper/2503.20641/) (2025)
- [Unlocking the Capabilities of Thought: A Reasoning Boundary Framework to Quantify and Optimize Chain-of-Thought](https://mlsanity.com/paper/2410.05695/) (2024)
- [Thought Manipulation: External Thought Can Be Efficient for Large Reasoning Models](https://mlsanity.com/paper/2504.13626/) (2025)

## Built on by (most cited first, 26 of 26)

- [Does Thinking More always Help? Mirage of Test-Time Scaling in Reasoning Models](https://mlsanity.com/paper/2506.04210/) (2025)
- [Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecycle](https://mlsanity.com/paper/2509.16679/) (2025)
- [Dynamic Model Routing and Cascading for Efficient LLM Inference: A Survey](https://mlsanity.com/paper/2603.04445/) (2026)
- [Towards Concise and Adaptive Thinking in Large Reasoning Models: A Survey](https://mlsanity.com/paper/2507.09662/) (2025)
- [CodeDance: A Dynamic Tool-integrated MLLM for Executable Visual Reasoning](https://mlsanity.com/paper/2512.17312/) (2025)
- [Revisiting Entropy Regularization: Adaptive Coefficient Unlocks Its Potential for LLM Reinforcement Learning](https://mlsanity.com/paper/2510.10959/) (2025)
- [Sparse-RL: Breaking the Memory Wall in LLM Reinforcement Learning via Stable Sparse Rollouts](https://mlsanity.com/paper/2601.10079/) (2026)
- [Hierarchical Budget Policy Optimization for Adaptive Reasoning](https://mlsanity.com/paper/2507.15844/) (2025)
- [Thinking Sparks!: Emergent Attention Heads in Reasoning Models During Post Training](https://mlsanity.com/paper/2509.25758/) (2025)
- [Test-Time Scaling in Diffusion LLMs via Hidden Semi-Autoregressive Experts](https://mlsanity.com/paper/2510.05040/) (2025)
- [Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models](https://mlsanity.com/paper/2601.11340/) (2026)
- [FoE: Forest of Errors Makes the First Solution the Best in Large Reasoning Models](https://mlsanity.com/paper/2604.02967/) (2026)
- [PaTaRM: Bridging Pairwise and Pointwise Signals via Preference-Aware Task-Adaptive Reward Modeling](https://mlsanity.com/paper/2510.24235/) (2025)
- [Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasoners](https://mlsanity.com/paper/2509.26226/) (2025)
- [DART: Difficulty-Adaptive Reasoning Truncation for Efficient Large Language Models](https://mlsanity.com/paper/2511.01170/) (2025)
- [Overthinking Reduction with Decoupled Rewards and Curriculum Data Scheduling](https://mlsanity.com/paper/2509.25827/) (2025)
- [Beyond Speedup -- Utilizing KV Cache for Sampling and Reasoning](https://mlsanity.com/paper/2601.20326/) (2026)
- [CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning](https://mlsanity.com/paper/2603.08659/) (2026)
- [Adaptive Dual Reasoner: Large Reasoning Models Can Think Efficiently by Hybrid Reasoning](https://mlsanity.com/paper/2510.10207/) (2025)
- [PAEC: Position-Aware Entropy Calibration for LLM Reasoning in RLVR](https://mlsanity.com/paper/2606.08543/) (2026)
- [VisRef: Visual Refocusing while Thinking Improves Test-Time Scaling in Multi-Modal Large Reasoning Models](https://mlsanity.com/paper/2603.00207/) (2026)
- [Balancing the Reasoning Load: Difficulty-Differentiated Policy Optimization with Length Redistribution for Efficient and Robust Reinforcement Learning](https://mlsanity.com/paper/2603.18533/) (2026)
- [Post-Reasoning: Improving the Performance of Non-Thinking Models at No Cost](https://mlsanity.com/paper/2605.06165/) (2026)
- [Cluster, Route, Escalate: Cascaded Framework for Cost-Aware LLM Serving](https://mlsanity.com/paper/2606.27457/) (2026)
- [ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling](https://mlsanity.com/paper/2608.10928/) (2026)
- [Construting Reverse Thinking: Developing Large Language Models' Reverse Thingking Ability](https://mlsanity.com/paper/2609.24760/) (2026)

## Abstract (arXiv)

Large reasoning models (LRMs) achieve remarkable performance via long reasoning chains, but often incur excessive computational overhead due to redundant reasoning, especially on simple tasks. In this work, we systematically quantify the upper bounds of LRMs under both Long-Thinking and No-Thinking modes, and uncover the phenomenon of "Internal Self-Recovery Mechanism" where models implicitly supplement reasoning during answer generation. Building on this insight, we propose Adaptive Self-Recovery Reasoning (ASRR), a framework that suppresses unnecessary reasoning and enables implicit recovery. By introducing accuracy-aware length reward regulation, ASRR adaptively allocates reasoning effort according to problem difficulty, achieving high efficiency with negligible performance sacrifice. Experiments across multiple benchmarks and models show that, compared with GRPO, ASRR reduces reasoning budget by up to 32.5% (1.5B) and 25.7% (7B) with minimal accuracy loss (1.2% and 0.6% pass@1), and significantly boosts harmless rates on safety benchmarks (up to +21.7%). Our results highlight the potential of ASRR for enabling efficient, adaptive, and safer reasoning in LRMs.