Image de couverture de livre

How Much Thinking is Enough? Quantifying and Understanding Redundancy in LLM Reasoning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhai, Zhiyuan, You, Xinkai, Yan, Wenjing, Wang, Xin
Format:	Preprint
Publié:	2026
Sujets:	Artificial Intelligence Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Think Just Enough: Sequence-Level Entropy as a Confidence Signal for LLM Reasoning
par: Sharma, Aman, et autres
Publié: (2025)

How Much Data Is Enough? Uniform Convergence Bounds for Generative & Vision-Language Models under Low-Dimensional Structure
par: Thompson, Paul M.
Publié: (2025)

Just Enough Thinking: Efficient Reasoning with Adaptive Length Penalties Reinforcement Learning
par: Xiang, Violet, et autres
Publié: (2025)

Think Clearly: Improving Reasoning via Redundant Token Pruning
par: Choi, Daewon, et autres
Publié: (2025)

Random Policy Valuation is Enough for LLM Reasoning with Verifiable Rewards
par: He, Haoran, et autres
Publié: (2025)

How Much Data is Enough? The Zeta Law of Discoverability in Biomedical Data, featuring the enigmatic Riemann zeta function
par: Thompson, Paul M.
Publié: (2026)

DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models
par: Shen, Yi, et autres
Publié: (2025)

Quantifying and Understanding Uncertainty in Large Reasoning Models
par: Li, Yangyi, et autres
Publié: (2026)

How Much Cache Does Reasoning Need? Depth-Cache Tradeoffs in KV-Compressed Transformers
par: Wang, Xiao
Publié: (2026)

How Much Backtracking is Enough? Exploring the Interplay of SFT and RL in Enhancing LLM Reasoning
par: Cai, Hongyi James, et autres
Publié: (2025)

Understanding Reasoning in Thinking Language Models via Steering Vectors
par: Venhoff, Constantin, et autres
Publié: (2025)

Infinite Width Models That Work: Why Feature Learning Doesn't Matter as Much as You Think
par: Sernau, Luke
Publié: (2024)

Accordion-Thinking: Self-Regulated Step Summaries for Efficient and Readable LLM Reasoning
par: Yang, Zhicheng, et autres
Publié: (2026)

Learn to Think: Bootstrapping LLM Reasoning Capability Through Graph Representation Learning
par: Gao, Hang, et autres
Publié: (2025)

Base Models Know How to Reason, Thinking Models Learn When
par: Venhoff, Constantin, et autres
Publié: (2025)

Is One Score Enough? Rethinking the Evaluation of Sequentially Evolving LLM Memory
par: Dong, Songwei, et autres
Publié: (2026)

Locality-Aware Redundancy Pruning for LLM Depth Compression
par: Yun, Vincent-Daniel, et autres
Publié: (2026)

DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models
par: Yin, Cheng, et autres
Publié: (2025)

Structural Reasoning Improves Molecular Understanding of LLM
par: Jang, Yunhui, et autres
Publié: (2024)

Is One Layer Enough? Understanding Inference Dynamics in Tabular Foundation Models
par: Balef, Amir Rezaei, et autres
Publié: (2026)

How Much is Too Much? Exploring LoRA Rank Trade-offs for Retaining Knowledge and Domain Robustness
par: Rathore, Darshita, et autres
Publié: (2025)

When, How Long and How Much? Interpretable Neural Networks for Time Series Regression by Learning to Mask and Aggregate
par: Forest, Florent, et autres
Publié: (2025)

Thinking with Deltas: Incentivizing Reinforcement Learning via Differential Visual Reasoning Policy
par: Gao, Shujian, et autres
Publié: (2026)

Comment on The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
par: Lawsen, A.
Publié: (2025)

ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spaces
par: Xu, Xin, et autres
Publié: (2026)

Think Before You Lie: How Reasoning Leads to Honesty
par: Yuan, Ann, et autres
Publié: (2026)

Do Large Language Models Know How Much They Know?
par: Prato, Gabriele, et autres
Publié: (2025)

How Much Can We Forget about Data Contamination?
par: Bordt, Sebastian, et autres
Publié: (2024)

How Many Experts Are Enough? Towards Optimal Semantic Specialization for Mixture-of-Experts
par: Park, Sumin, et autres
Publié: (2025)

AdaReasoner: Adaptive Reasoning Enables More Flexible Thinking in Large Language Models
par: Wang, Xiangqi, et autres
Publié: (2025)

Textual Bayes: Quantifying Prompt Uncertainty in LLM-Based Systems
par: Ross, Brendan Leigh, et autres
Publié: (2025)

Scaling-Aware Adapter for Structure-Grounded LLM Reasoning
par: Jing, Zihao, et autres
Publié: (2026)

Think-Augmented Function Calling: Improving LLM Parameter Accuracy Through Embedded Reasoning
par: Wei, Lei, et autres
Publié: (2026)

Towards Mitigation of Hallucination for LLM-empowered Agents: Progressive Generalization Bound Exploration and Watchdog Monitor
par: Liu, Siyuan, et autres
Publié: (2025)

Efficient Reasoning with Hidden Thinking
par: Shen, Xuan, et autres
Publié: (2025)

One Token Embedding Is Enough to Deadlock Your Large Reasoning Model
par: Zhang, Mohan, et autres
Publié: (2025)

Flow-of-Options: Diversified and Improved LLM Reasoning by Thinking Through Options
par: Nair, Lakshmi, et autres
Publié: (2025)

Dynamic Thinking-Token Selection for Efficient Reasoning in Large Reasoning Models
par: Guo, Zhenyuan, et autres
Publié: (2026)

Think in Blocks: Adaptive Reasoning from Direct Response to Deep Reasoning
par: Zhu, Yekun, et autres
Publié: (2025)

How Much Progress Did I Make? An Unexplored Human Feedback Signal for Teaching Robots
par: Yu, Hang, et autres
Publié: (2024)