Image de couverture de livre

On the Overscaling Curse of Parallel Thinking: System Efficacy Contradicts Sample Efficiency

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Wang, Yiming, Zhang, Zhuosheng, Wang, Rui
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning Artificial Intelligence 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

Embedding Trajectory for Out-of-Distribution Detection in Mathematical Reasoning
par: Wang, Yiming, et autres
Publié: (2024)

An Analysis and Mitigation of the Reversal Curse
par: Lv, Ang, et autres
Publié: (2023)

Mitigating Reversal Curse in Large Language Models via Semantic-aware Permutation Training
par: Guo, Qingyan, et autres
Publié: (2024)

The Factorization Curse: Which Tokens You Predict Underlie the Reversal Curse and More
par: Kitouni, Ouail, et autres
Publié: (2024)

Demystifying Hybrid Thinking: Can LLMs Truly Switch Between Think and No-Think?
par: Wang, Shouren, et autres
Publié: (2025)

Exploring the Reversal Curse and Other Deductive Logical Reasoning in BERT and GPT-Based Large Language Models
par: Wu, Da, et autres
Publié: (2023)

Sampling-Efficient Test-Time Scaling: Self-Estimating the Best-of-N Sampling in Early Decoding
par: Wang, Yiming, et autres
Publié: (2025)

To Think or Not to Think: Exploring the Unthinking Vulnerability in Large Reasoning Models
par: Zhu, Zihao, et autres
Publié: (2025)

Latent Space Chain-of-Embedding Enables Output-free LLM Self-Evaluation
par: Wang, Yiming, et autres
Publié: (2024)

Do Depth-Grown Models Overcome the Curse of Depth? An In-Depth Analysis
par: Kapl, Ferdinand, et autres
Publié: (2025)

MoEs Are Stronger than You Think: Hyper-Parallel Inference Scaling with RoE
par: Zibakhsh, Soheil, et autres
Publié: (2025)

Thinking About Thinking: SAGE-nano's Inverse Reasoning for Self-Aware Language Models
par: Jha, Basab, et autres
Publié: (2025)

The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
par: Berglund, Lukas, et autres
Publié: (2023)

StableSSM: Alleviating the Curse of Memory in State-space Models through Stable Reparameterization
par: Wang, Shida, et autres
Publié: (2023)

Optimizing Temperature for Language Models with Multi-Sample Inference
par: Du, Weihua, et autres
Publié: (2025)

WikiContradict: A Benchmark for Evaluating LLMs on Real-World Knowledge Conflicts from Wikipedia
par: Hou, Yufang, et autres
Publié: (2024)

Reasoning with Sampling: Your Base Model is Smarter Than You Think
par: Karan, Aayush, et autres
Publié: (2025)

Thoughtbubbles: an Unsupervised Method for Parallel Thinking in Latent Space
par: Liu, Houjun, et autres
Publié: (2025)

AdapThink: Adaptive Thinking Preferences for Reasoning Language Model
par: Wan, Xu, et autres
Publié: (2025)

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

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

Learning to Refine: Self-Refinement of Parallel Reasoning in LLMs
par: Wang, Qibin, et autres
Publié: (2025)

AdaptThink: Reasoning Models Can Learn When to Think
par: Zhang, Jiajie, et autres
Publié: (2025)

More Thinking, More Bias: Length-Driven Position Bias in Reasoning Models
par: Wang, Xiao
Publié: (2026)

Efficient Reasoning with Balanced Thinking
par: Li, Yulin, et autres
Publié: (2026)

Process Reward Models That Think
par: Khalifa, Muhammad, et autres
Publié: (2025)

MeTHanol: Modularized Thinking Language Models with Intermediate Layer Thinking, Decoding and Bootstrapping Reasoning
par: Xi, Ningyuan, et autres
Publié: (2024)

FlashSampling: Fast and Memory-Efficient Exact Sampling
par: Ruiz, Tomas, et autres
Publié: (2026)

CoRT: Code-integrated Reasoning within Thinking
par: Li, Chengpeng, et autres
Publié: (2025)

Effectively Controlling Reasoning Models through Thinking Intervention
par: Wu, Tong, et autres
Publié: (2025)

Think Before You Prune: Self-Reflective Structured Pruning for Reasoning Language Models
par: Wang, Ziyan, et autres
Publié: (2025)

BRiTE: Bootstrapping Reinforced Thinking Process to Enhance Language Model Reasoning
par: Zhong, Han, et autres
Publié: (2025)

Adaptive Milestone Reward for GUI Agents
par: Zheng, Congmin, et autres
Publié: (2026)

Training Large Language Models To Reason In Parallel With Global Forking Tokens
par: Jia, Sheng, et autres
Publié: (2025)

Parallel Test-Time Scaling for Latent Reasoning Models
par: You, Runyang, et autres
Publié: (2025)

TingIS: Real-time Risk Event Discovery from Noisy Customer Incidents at Enterprise Scale
par: Wang, Jun, et autres
Publié: (2026)

From Faithfulness to Correctness: Generative Reward Models that Think Critically
par: Ma, Qiyao, et autres
Publié: (2025)

SenTSR-Bench: Thinking with Injected Knowledge for Time-Series Reasoning
par: He, Zelin, et autres
Publié: (2026)

Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggers
par: Yang, Wang, et autres
Publié: (2026)

To Think or Not to Think: The Hidden Cost of Meta-Training with Excessive CoT Examples
par: Kothapalli, Vignesh, et autres
Publié: (2025)