mlsanityData to 14 Oct 2026

Deep Think with Confidence

arXiv 2508.15260 ↗First posted 21 Aug 2025cs.LGcited by 205

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

Abstract (arXiv)

Large Language Models (LLMs) have shown great potential in reasoning tasks through test-time scaling methods like self-consistency with majority voting. However, this approach often leads to diminishing returns in accuracy and high computational overhead. To address these challenges, we introduce Deep Think with Confidence (DeepConf), a simple yet powerful method that enhances both reasoning efficiency and performance at test time. DeepConf leverages model-internal confidence signals to dynamically filter out low-quality reasoning traces during or after generation. It requires no additional model training or hyperparameter tuning and can be seamlessly integrated into existing serving frameworks. We evaluate DeepConf across a variety of reasoning tasks and the latest open-source models, including Qwen 3 and GPT-OSS series. Notably, on challenging benchmarks such as AIME 2025, DeepConf@512 achieves up to 99.9% accuracy and reduces generated tokens by up to 84.7% compared to full parallel thinking.

Key facts. Cited by 205 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 34 of its 43 references that are in the map. First posted on arXiv 21 Aug 2025.

Builds on

34 of 43 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025Don't Overthink it. Preferring Shorter Thinking Chains for Improved LLM Reasoning23 May
  2. 2025Ranked Voting based Self-Consistency of Large Language Models16 May
  3. 2025Dynamic Early Exit in Reasoning Models22 Apr
  4. 2025ThinkPrune: Pruning Long Chain-of-Thought of LLMs via Reinforcement Learning2 Apr
  5. 2025Scalable Best-of-N Selection for Large Language Models via Self-Certainty25 Feb
  6. 2025s1: Simple test-time scaling31 Jan
  7. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
  8. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  9. 2024Efficiently Scaling LLM Reasoning with Certaindex30 Dec
  10. 2024Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems12 Dec
  11. 2024Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation3 Oct
  12. 2024Reasoning Aware Self-Consistency: Leveraging Reasoning Paths for Efficient LLM Sampling30 Aug
4 more
  1. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug
  2. 2024Soft Self-Consistency Improves Language Model Agents20 Feb
  3. 2024Escape Sky-high Cost: Early-stopping Self-Consistency for Multi-step Reasoning19 Jan
  4. 2023Let's Sample Step by Step: Adaptive-Consistency for Efficient Reasoning and Coding with LLMs19 May

Other references · 18

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  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. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 991
  8. 2025LIMO: Less is More for Reasoningcited by 498
  9. 2025MathArena: Evaluating LLMs on Uncontaminated Math Competitionscited by 297
  10. 2023A Survey of Confidence Estimation and Calibration in Large Language Modelscited by 256
  11. 2025Learning to Reason without External Rewardscited by 210
  12. 2024Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantificationcited by 162
6 more
  1. 2024From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Modelscited by 137
  2. 2023Self-Evaluation Improves Selective Generation in Large Language Modelscited by 76
  3. 2024Learning to Route LLMs with Confidence Tokenscited by 67
  4. 2023Rewarding Chatbots for Real-World Engagement with Millions of Userscited by 59
  5. 2023Lightweight reranking for language model generationscited by 10
  6. 2024Not All Votes Count! Programs as Verifiers Improve Self-Consistency of Language Models for Math Reasoningcited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 73; 2026: 132.

205 papers.

Sort:

By area:

  1. 2025TTT3R: 3D Reconstruction as Test-Time Trainingcited by 129
  2. 2025Parallel-R1: Towards Parallel Thinking via Reinforcement Learningcited by 61
  3. 2025Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Modelscited by 48
  4. 2025The Majority is not always right: RL training for solution aggregationcited by 47
  5. 2025Reverse-Engineered Reasoning for Open-Ended Generationcited by 32
  6. 2025Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoningcited by 31
  7. 2025CDE: Curiosity-Driven Exploration for Efficient Reinforcement Learning in Large Language Modelscited by 29
  8. 2025Single-stream Policy Optimizationcited by 28
  9. 2026Adaptive Action Chunking at Inference-time for Vision-Language-Action Modelscited by 28
  10. 2025Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Rewardcited by 27
  11. 2025Beyond Majority Voting: LLM Aggregation by Leveraging Higher-Order Informationcited by 26
  12. 2026Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokenscited by 22
38 more
  1. 2025Rethinking Thinking Tokens: LLMs as Improvement Operatorscited by 21
  2. 2026Uncertainty Quantification in LLM Agents: Foundations, Emerging Challenges, and Opportunitiescited by 20
  3. 2025Trace Length is a Simple Uncertainty Signal in Reasoning Modelscited by 19
  4. 2025Mitigating LLM Hallucination via Behaviorally Calibrated Reinforcement Learningcited by 18
  5. 2026Agentic Aggregation for Parallel Scaling of Long-Horizon Agentic Taskscited by 18
  6. 2026Equilibrium Reasoners: Learning Attractors Enables Scalable Reasoningcited by 18
  7. 2025Stop When Enough: Adaptive Early-Stopping for Chain-of-Thought Reasoningcited by 16
  8. 2025TUMIX: Multi-Agent Test-Time Scaling with Tool-Use Mixturecited by 15
  9. 2025Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thoughtcited by 15
  10. 2026Agentic Confidence Calibrationcited by 15
  11. 2025Parallel Test-Time Scaling for Latent Reasoning Modelscited by 14
  12. 2025ParallelMuse: Agentic Parallel Thinking for Deep Information Seekingcited by 14
  13. 2026Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Modelscited by 14
  14. 2026Parallel-Probe: Towards Efficient Parallel Thinking via 2D Probingcited by 14
  15. 2025Energy Use of AI Inference, Efficiency Pathways, and Test-Time Scalingcited by 13
  16. 2026Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scalingcited by 13
  17. 2025CLUE: Non-parametric Verification from Experience via Hidden-State Clusteringcited by 12
  18. 2025Certified Self-Consistency: Statistical Guarantees and Test-Time Training for Reliable Reasoning in LLMscited by 12
  19. 2025From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Modelscited by 12
  20. 2025A Survey on Parallel Reasoningcited by 11
  21. 2026Reliability-Aware Adaptive Self-Consistency for Efficient Sampling in LLM Reasoningcited by 11
  22. 2025Self-Evolving Vision-Language Models for Image Quality Assessment via Voting and Rankingcited by 10
  23. 2025Let it Calm: Exploratory Annealed Decoding for Verifiable Reinforcement Learningcited by 10
  24. 2025EAGer: Entropy-Aware GEneRation for Adaptive Inference-Time Scalingcited by 10
  25. 2025Scaling Test-Time Compute to Achieve IOI Gold Medal with Open-Weight Modelscited by 10
  26. 2025ReProbe: Efficient Test-Time Scaling of Multi-Step Reasoning by Probing Internal States of Large Language Modelscited by 10
  27. 2026Hidden States as Early Signals: Step-level Trace Evaluation and Pruning for Efficient Test-Time Scalingcited by 10
  28. 2026Multi-Agent Debate with Memory Maskingcited by 10
  29. 2026LLMs Improving LLMs: Agentic Discovery for Test-Time Scalingcited by 10
  30. 2025Klear-AgentForge: Forging Agentic Intelligence through Posttraining Scalingcited by 9
  31. 2026Do Not Waste Your Rollouts: Recycling Search Experience for Efficient Test-Time Scalingcited by 9
  32. 2025ThreadWeaver: Adaptive Threading for Efficient Parallel Reasoning in Language Modelscited by 8
  33. 2026ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spacescited by 8
  34. 2026Agentic Test-Time Scaling for WebAgentscited by 8
  35. 2026Reported Confidence in LLMs Tracks Commitment More Than Correctnesscited by 8
  36. 2025Selective Expert Guidance for Effective and Diverse Exploration in Reinforcement Learning of LLMscited by 7
  37. 2025Thinking on the Fly: Test-Time Reasoning Enhancement via Latent Thought Policy Optimizationcited by 7
  38. 2025Mixture-of-Minds: Multi-Agent Reinforcement Learning for Table Understandingcited by 7

50 most cited of 205 All 205 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025DeepDistill: Enhancing LLM Reasoning Capabilities via Large-Scale Difficulty-Graded Data Training
  2. 2025ConCISE: Confidence-guided Compression in Step-by-step Efficient Reasoning
  3. 2026Beyond Model Scaling: Test-Time Intervention for Efficient Deep Reasoning
  4. 2026ThinkBooster: A Unified Framework for Seamless Test-Time Scaling of LLM Reasoning
  5. 2025Learning to Reason: Training LLMs with GPT-OSS or DeepSeek R1 Reasoning Traces
  6. 2025Benchmarking Reasoning Robustness in Large Language Models
  7. 2025DeepQuestion: Systematic Generation of Real-World Challenges for Evaluating LLMs Performance
  8. 2026Confidence over Time: Confidence Calibration with Temporal Logic for Large Language Model Reasoning