Image de couverture de livre

Step Potential Advantage Estimation: Harnessing Intermediate Confidence and Correctness for Efficient Mathematical Reasoning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Wu, Fei, Zhang, Zhenrong, Chang, Qikai, Zhang, Jianshu, Liu, Quan, Du, Jun
Format:	Preprint
Publié:	2026
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

THOR: Tool-Integrated Hierarchical Optimization via RL for Mathematical Reasoning
par: Chang, Qikai, et autres
Publié: (2025)

PEARL: Training Socratic Tutors with Pedagogically Aligned Reinforcement Learning
par: Chang, Qikai, et autres
Publié: (2026)

DocMamba: Efficient Document Pre-training with State Space Model
par: Hu, Pengfei, et autres
Publié: (2024)

Stabilizing Efficient Reasoning with Step-Level Advantage Selection
par: Wang, Han, et autres
Publié: (2026)

DuaShepherd: Integrating Stepwise Correctness and Potential Rewards for Mathematical Reasoning
par: Wu, Yuanhao, et autres
Publié: (2025)

KTAE: A Model-Free Algorithm to Key-Tokens Advantage Estimation in Mathematical Reasoning
par: Sun, Wei, et autres
Publié: (2025)

Enhancing the Geometric Problem-Solving Ability of Multimodal LLMs via Symbolic-Neural Integration
par: Pan, Yicheng, et autres
Publié: (2025)

Enhancing Mathematical Reasoning in LLMs by Stepwise Correction
par: Wu, Zhenyu, et autres
Publié: (2024)

Skeleton and Font Generation Network for Zero-shot Chinese Character Generation
par: Xue, Mobai, et autres
Publié: (2025)

SRFUND: A Multi-Granularity Hierarchical Structure Reconstruction Benchmark in Form Understanding
par: Ma, Jiefeng, et autres
Publié: (2024)

AdaptiveStep: Automatically Dividing Reasoning Step through Model Confidence
par: Liu, Yuliang, et autres
Publié: (2025)

ConCISE: Confidence-guided Compression in Step-by-step Efficient Reasoning
par: Qiao, Ziqing, et autres
Publié: (2025)

First-Step Advantage: Importance of Starting Right in Multi-Step Math Reasoning
par: Jain, Kushal, et autres
Publié: (2023)

Full-Step-DPO: Self-Supervised Preference Optimization with Step-wise Rewards for Mathematical Reasoning
par: Xu, Huimin, et autres
Publié: (2025)

ConfSpec: Efficient Step-Level Speculative Reasoning via Confidence-Gated Verification
par: Liu, Siran, et autres
Publié: (2026)

Step Guided Reasoning: Improving Mathematical Reasoning using Guidance Generation and Step Reasoning
par: Cao, Lang, et autres
Publié: (2024)

Let's Learn Step by Step: Enhancing In-Context Learning Ability with Curriculum Learning
par: Liu, Yinpeng, et autres
Publié: (2024)

Soundness-Aware Level: A Microscopic Signature that Predicts LLM Reasoning Potential
par: Wu, Xuansheng, et autres
Publié: (2025)

Supervised Optimism Correction: Be Confident When LLMs Are Sure
par: Zhang, Junjie, et autres
Publié: (2025)

All Roads Lead to Rome: Graph-Based Confidence Estimation for Large Language Model Reasoning
par: Zhang, Caiqi, et autres
Publié: (2025)

Fact-and-Reflection (FaR) Improves Confidence Calibration of Large Language Models
par: Zhao, Xinran, et autres
Publié: (2024)

MMBoundary: Advancing MLLM Knowledge Boundary Awareness through Reasoning Step Confidence Calibration
par: He, Zhitao, et autres
Publié: (2025)

Evaluating Intermediate Reasoning of Code-Assisted Large Language Models for Mathematics
par: Al-Khalili, Zena, et autres
Publié: (2025)

Curing Miracle Steps in LLM Mathematical Reasoning with Rubric Rewards
par: Yuan, Youliang, et autres
Publié: (2025)

SMRC: Aligning Large Language Models with Student Reasoning for Mathematical Error Correction
par: Zeng, Biaojie, et autres
Publié: (2025)

SHAPE: Stage-aware Hierarchical Advantage via Potential Estimation for LLM Reasoning
par: Ai, Zhengyang, et autres
Publié: (2026)

MathFimer: Enhancing Mathematical Reasoning by Expanding Reasoning Steps through Fill-in-the-Middle Task
par: Yan, Yuchen, et autres
Publié: (2025)

Outcome-Grounded Advantage Reshaping for Fine-Grained Credit Assignment in Mathematical Reasoning
par: Li, Ziheng, et autres
Publié: (2026)

Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation
par: Li, Changcheng, et autres
Publié: (2026)

PolyMath: Evaluating Mathematical Reasoning in Multilingual Contexts
par: Wang, Yiming, et autres
Publié: (2025)

Conjecturing: An Overlooked Step in Formal Mathematical Reasoning
par: Sivakumar, Jasivan Alex, et autres
Publié: (2025)

Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning
par: Zhao, Zhengyang, et autres
Publié: (2026)

PRM-BAS: Enhancing Multimodal Reasoning through PRM-guided Beam Annealing Search
par: Hu, Pengfei, et autres
Publié: (2025)

An Efficient and Precise Training Data Construction Framework for Process-supervised Reward Model in Mathematical Reasoning
par: Sun, Wei, et autres
Publié: (2025)

Conditional Advantage Estimation for Reinforcement Learning in Large Reasoning Models
par: Chen, Guanxu, et autres
Publié: (2025)

EquivPruner: Boosting Efficiency and Quality in LLM-Based Search via Action Pruning
par: Liu, Jiawei, et autres
Publié: (2025)

Confidence Geometry Reveals Trace-Level Correctness in Large Language Model Reasoning
par: Liu, Shuo, et autres
Publié: (2026)

Evaluating Step-by-Step Reasoning through Symbolic Verification
par: Zhang, Yi-Fan, et autres
Publié: (2022)

Confidence-Calibrated Small-Large Language Model Collaboration for Cost-Efficient Reasoning
par: Zhang, Chuang, et autres
Publié: (2026)

LogicReward: Incentivizing LLM Reasoning via Step-Wise Logical Supervision
par: Xu, Jundong, et autres
Publié: (2025)