mlsanityData to 7 Oct 2026

Step-level Verifier-guided Hybrid Test-Time Scaling for Large Language Models

arXiv 2507.15512 ↗First posted 21 Jul 2025cs.CLcited by 10

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

Abstract (arXiv)

Test-Time Scaling (TTS) is a promising approach to progressively elicit the model's intelligence during inference. Recently, training-based TTS methods, such as continued reinforcement learning (RL), have further surged in popularity, while training-free TTS methods are gradually fading from prominence. However, the additional computation overhead of training amplifies the burden on test-time scaling. In this paper, we focus on training-free TTS methods for reasoning. We first design Conditional Step-level Self-refinement, a fine-grained sequential scaling method guided by process verification. On top of its effectiveness, we further combine it with other classical parallel scaling methods at the step level, to introduce a novel inference paradigm called Hybrid Test-Time Scaling. Extensive experiments on five instruction-tuned LLMs across different scales (3B-14B) and families demonstrate that hybrid strategy incorporating various training-free TTS methods at a fine granularity has considerable potential for expanding the reasoning performance boundaries of LLMs.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 28 of its 31 references that are in the map. First posted on arXiv 21 Jul 2025.

Builds on

28 of 31 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling10 Feb
  2. 2025s1: Simple test-time scaling31 Jan
  3. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  4. 2024Marco-o1: Towards Open Reasoning Models for Open-Ended Solutions21 Nov
  5. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug
  6. 2024Small Language Models Need Strong Verifiers to Self-Correct Reasoning26 Apr
  7. 2023Large Language Models Cannot Self-Correct Reasoning Yet3 Oct
  8. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing19 May
  9. 2023Self-Refine: Iterative Refinement with Self-Feedback30 Mar

Other references · 19

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  2. 2020Scaling Laws for Neural Language Modelscited by 6,411
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  6. 2023Let's Verify Step by Stepcited by 3,598
  7. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  8. 2019Mastering Atari, Go, Chess and Shogi by Planning with a Learned Modelcited by 1,327
  9. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 946
  10. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 895
  11. 2022Solving math word problems with process- and outcome-based feedbackcited by 871
  12. 2024Generative Verifiers: Reward Modeling as Next-Token Predictioncited by 465
7 more
  1. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 402
  2. 2024Accessing GPT-4 level Mathematical Olympiad Solutions via Monte Carlo Tree Self-refine with LLaMa-3 8Bcited by 156
  3. 2025PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Modelscited by 98
  4. 2024OpenR: An Open Source Framework for Advanced Reasoning with Large Language Modelscited by 92
  5. 2024Enhancing Reasoning Capabilities of LLMs via Principled Synthetic Logic Corpuscited by 29
  6. 2025GRAM: A Generative Foundation Reward Model for Reward Generalizationcited by 25
  7. 2025SLAM: Towards Efficient Multilingual Reasoning via Selective Language Alignmentcited by 17

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 9.

10 papers.

  1. 2026SWE-TRACE: Optimizing Long-Horizon SWE Agents Through Rubric Process Reward Models and Heuristic Test-Time Scalingcited by 12
  2. 2026Tool Verification for Test-Time Reinforcement Learningcited by 6
  3. 2025AlphaApollo: A System for Deep Agentic Reasoningcited by 5
  4. 2026MSRL: Scaling Generative Multimodal Reward Modeling via Multi-Stage Reinforcement Learningcited by 5
  5. 2026Distribution-Aware Reward Estimation for Test-Time Reinforcement Learningcited by 2
  6. 2026What Am I Missing? Question-Answering as Hidden State Probingcited by 1
  7. 2026APR: Penalizing Structural Redundancy in Large Reasoning Models via Anchor-based Process Rewardscited by 0
  8. 2026Test-Time Scaling in the Wild: Why Exploitation, Not Exploration, Is the Bottleneckcited by 0
  9. 2026Recursive Agentic Reasoningcited by 0
  10. 2026Beyond the Answer Key: Robustness Evaluation of Large Language Models for Step-Level Mathematical Verificationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Rethinking Inference-Time Scaling: Efficiency Limits and Linguistic Signals
  2. 2025Are We Scaling the Right Thing? A System Perspective on Test-Time Scaling
  3. 2025The Art of Scaling Test-Time Compute for Large Language Models
  4. 2025Trust but Verify! A Survey on Verification Design for Test-time Scaling
  5. 2025T1: Advancing Language Model Reasoning through Reinforcement Learning and Inference Scaling
  6. 2025On the Role of Temperature Sampling in Test-Time Scaling
  7. 2026ThinkBooster: A Unified Framework for Seamless Test-Time Scaling of LLM Reasoning
  8. 2025Prompting Test-Time Scaling Is A Strong LLM Reasoning Data Augmentation