mlsanityData to 14 Oct 2026

ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling

arXiv 2608.10928 ↗First posted 11 Aug 2026cs.AIcited by 0

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

also in 1 more
Abstract (arXiv)

Large Reasoning Models (LRMs) improve performance by allocating additional inference-time compute to generate extended chain-of-thought reasoning. However, recent studies reveal that sequential test-time scaling often yields diminishing or even negative returns, as longer traces exhibit increased uncertainty, error compounding, and drift from the original problem. We propose ThinkRetrieve, a test-time scaling framework that augments the reasoning traces of LRMs with dynamically retrieved solved examples at each reasoning step. Given an external corpus of problems paired with step-by-step solutions, ThinkRetrieve retrieves relevant exemplars at each intermediate step and injects them directly into the thinking trace, providing the model with guidance on how to reason rather than merely what facts are relevant. Experiments across five reasoning models (1.5B--8B parameters) on GSM-8K, MATH-500, AIME 2025, and SciQ demonstrate that ThinkRetrieve consistently improves accuracy over standard test-time scaling, with relative gains of up to 60% on AIME 2025.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 44 of its 65 references that are in the map. First posted on arXiv 11 Aug 2026.

Builds on

44 of 65 references.

Closest ancestors · 15

Same direction, newest first

  1. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Models4 Aug
  2. 2025Inverse Scaling in Test-Time Compute19 Jul
  3. 2025The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity7 Jun
  4. 2025AdaCtrl: Towards Adaptive and Controllable Reasoning via Difficulty-Aware Budgeting24 May
  5. 2025When to Continue Thinking: Adaptive Thinking Mode Switching for Efficient Reasoning21 May
  6. 2025Think Only When You Need with Large Hybrid-Reasoning Models20 May
  7. 2025ThinkSwitcher: When to Think Hard, When to Think Fast20 May
  8. 2025AdaptThink: Reasoning Models Can Learn When to Think19 May
  9. 2025Thinkless: LLM Learns When to Think19 May
  10. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning6 Mar
  11. 2025Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning25 Feb
  12. 2025s1: Simple test-time scaling31 Jan
3 more
  1. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  2. 2024O1 Replication Journey: A Strategic Progress Report -- Part 18 Oct
  3. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug

Other references · 29

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  4. 2025Qwen3 Technical Reportcited by 8,820
  5. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  6. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  7. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  8. 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,212
  9. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,444
  10. 2022Text Embeddings by Weakly-Supervised Contrastive Pre-trainingcited by 1,253
  11. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  12. 2024DeepSeek-V3 Technical Reportcited by 811
17 more
  1. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Modelscited by 665
  2. 2024GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Modelscited by 550
  3. 2024BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructionscited by 546
  4. 2025VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Modelcited by 526
  5. 2025Visual-RFT: Visual Reinforcement Fine-Tuningcited by 513
  6. 2025Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Modelscited by 293
  7. 2025Training Language Models to Reason Efficientlycited by 247
  8. 2024OpenR: An Open Source Framework for Advanced Reasoning with Large Language Modelscited by 94
  9. 2024LLaMA-Berry: Pairwise Optimization for O1-like Olympiad-Level Mathematical Reasoningcited by 91
  10. 2024RAT: Retrieval Augmented Thoughts Elicit Context-Aware Reasoning in Long-Horizon Generationcited by 75
  11. 2023Dr.ICL: Demonstration-Retrieved In-context Learningcited by 72
  12. 2025Retrieval-Augmented Process Reward Model for Generalizable Mathematical Reasoningcited by 23
  13. 2025REAL-Prover: Retrieval Augmented Lean Prover for Mathematical Reasoningcited by 18
  14. 2025RARE: Retrieval-Augmented Reasoning Modelingcited by 11
  15. 2025Reasoning Large Language Model Errors Arise from Hallucinating Critical Problem Featurescited by 6
  16. 2026MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrievalcited by 6
  17. 2025Guideline Forest: Retrieval-Augmented Reasoning with Branching Experience-Induced Guidelinescited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Thought calibration: Efficient and confident test-time scaling
  2. 2025Scaling over Scaling: Exploring Test-Time Scaling Plateau in Large Reasoning Models
  3. 2025The Art of Scaling Test-Time Compute for Large Language Models
  4. 2026Thinking Long, but Short: Stable Sequential Test-Time Scaling for Large Reasoning Models
  5. 2026Efficient Test-Time Scaling via Temporal Reasoning Aggregation
  6. 2025ARISE: An Adaptive Resolution-Aware Metric for Test-Time Scaling Evaluation in Large Reasoning Models
  7. 2025Retrieval-of-Thought: Efficient Reasoning via Reusing Thoughts
  8. 2025THINK-Bench: Evaluating Thinking Efficiency and Chain-of-Thought Quality of Large Reasoning Models