mlsanityData to 14 Oct 2026

Stop When Further Reasoning Won't Help: Attention-State Adaptive Generation in Reasoning Models

arXiv 2606.15070 ↗First posted 13 Jun 2026cs.CLcited by 3

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

Abstract (arXiv)

By incorporating test-time compute scaling, large reasoning models (LRMs) can solve complex problems through explicit chain-of-thought (CoT) reasoning processes. However, they often suffer from overthinking, resulting in redundant token outputs and degraded accuracy. Current methods to mitigate this issue remain limited: training-based approaches require substantial computational resources, while training-free methods rely on well-crafted prompts or unreliable confidence signals. In this work, we investigate early stopping from the perspective of attention distributions and propose a simple method, ASAG, which infers the model's reasoning state and adaptively adjusts the generation strategy. The proposed framework is training-free and plug-and-play, enabling seamless integration into existing LRMs. Extensive experiments on nine benchmarks demonstrate consistent improvements across mainstream LRMs with varying parameter scales, including the DeepSeek-R1-Distill and Qwen3 series. Specifically, ASAG improves average accuracy by 3.2% while reducing the number of generated tokens by nearly 40% across all reasoning tasks on Qwen3-8B.

Key facts. Cited by 3 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 19 of its 48 references that are in the map. First posted on arXiv 13 Jun 2026.

Builds on

19 of 48 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025Do Thinking Tokens Help or Trap? Towards More Efficient Large Reasoning Model30 Jun
  2. 2025Dynamic Early Exit in Reasoning Models22 Apr
  3. 2025Reasoning Models Can Be Effective Without Thinking14 Apr
  4. 2024Efficiently Scaling LLM Reasoning with Certaindex30 Dec
  5. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug

Other references · 14

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  5. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  6. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  7. 2025LIMO: Less is More for Reasoningcited by 498
  8. 2025Humanity's Last Examcited by 491
  9. 2025Chain of Draft: Thinking Faster by Writing Lesscited by 297
  10. 2024Distilling System 2 into System 1cited by 159
  11. 2025Deconstructing Long Chain-of-Thought: A Structured Reasoning Optimization Framework for Long CoT Distillationcited by 43
  12. 2026Why Steering Works: Toward a Unified View of Language Model Parameter Dynamicscited by 9
2 more
  1. 2026CBV: Clean-label Backdoor Attacks on Vision Language Models via Diffusion Modelscited by 3
  2. 2026Toward Robust GraphRAG: Mitigating Retrieval Drift and Hallucination from Imperfect Knowledge Graphscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controllercited by 0
  2. 2026S3: Spectral Null-Space Swap Makes Reasoning Models Efficientcited by 0
  3. 2026RIVO: Lightweight Adaptive Rerouting for Deep Research Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Improving Chain-of-Thought for Logical Reasoning via Attention-Aware Intervention
  2. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models
  3. 2025Prompting Test-Time Scaling Is A Strong LLM Reasoning Data Augmentation
  4. 2026TERMINATOR: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoning
  5. 2025From Long to Short: LLMs Excel at Trimming Own Reasoning Chains
  6. 2025A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond
  7. 2026Thinking Long, but Short: Stable Sequential Test-Time Scaling for Large Reasoning Models
  8. 2025Encode, Think, Decode: Scaling test-time reasoning with recursive latent thoughts