Cover Image

Between Underthinking and Overthinking: An Empirical Study of Reasoning Length and correctness in LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Su, Jinyan, Healey, Jennifer, Nakov, Preslav, Cardie, Claire
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Adapting Fake News Detection to the Era of Large Language Models
by: Su, Jinyan, et al.
Published: (2023)

Fast or Better? Balancing Accuracy and Cost in Retrieval-Augmented Generation with Flexible User Control
by: Su, Jinyan, et al.
Published: (2025)

Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewards
by: Su, Jinyan, et al.
Published: (2025)

Reasoning Court: Combining Reasoning, Action, and Judgment for Multi-Hop Reasoning
by: Wu, Jingtian, et al.
Published: (2025)

Knowing but Not Showing: LLMs Recognize Ambiguity but Rarely Ask Clarifying Questions
by: Su, Jinyan, et al.
Published: (2026)

Corpus Poisoning via Approximate Greedy Gradient Descent
by: Su, Jinyan, et al.
Published: (2024)

Multi-Hop Question Answering: When Can Humans Help, and Where do They Struggle?
by: Su, Jinyan, et al.
Published: (2025)

Avoiding Overthinking and Underthinking: Curriculum-Aware Budget Scheduling for LLMs
by: Rahman, Amirul, et al.
Published: (2026)

Cross-Cultural Transfer of Commonsense Reasoning in LLMs: Evidence from the Arab World
by: Almheiri, Saeed, et al.
Published: (2025)

From Chaos to Clarity: Claim Normalization to Empower Fact-Checking
by: Sundriyal, Megha, et al.
Published: (2023)

HAPO: Training Language Models to Reason Concisely via History-Aware Policy Optimization
by: Huang, Chengyu, et al.
Published: (2025)

Mitigating Overthinking through Reasoning Shaping
by: Song, Feifan, et al.
Published: (2025)

Multimodal Large Language Models to Support Real-World Fact-Checking
by: Geng, Jiahui, et al.
Published: (2024)

Better LLM Reasoning via Dual-Play
by: Zhang, Zhengxin, et al.
Published: (2025)

Profiling News Media for Factuality and Bias Using LLMs and the Fact-Checking Methodology of Human Experts
by: Mujahid, Zain Muhammad, et al.
Published: (2025)

Clarification Is Not Enough: Post-Clarification Answering Remains the Bottleneck in Multi-Turn QA
by: Su, Jinyan, et al.
Published: (2026)

COMMUNITYNOTES: A Dataset for Exploring the Helpfulness of Fact-Checking Explanations
by: Xing, Rui, et al.
Published: (2025)

HALF: Harm-Aware LLM Fairness Evaluation Aligned with Deployment
by: Mekky, Ali, et al.
Published: (2025)

SmartSwitch: Advancing LLM Reasoning by Overcoming Underthinking via Promoting Deeper Thought Exploration
by: Zhang, Xichen, et al.
Published: (2025)

Detecting Propaganda Techniques in Code-Switched Social Media Text
by: Salman, Muhammad Umar, et al.
Published: (2023)

Mitigating Overthinking in Large Reasoning Language Models via Reasoning Path Deviation Monitoring
by: Guan, Weixin, et al.
Published: (2026)

The Evolution of Thought: Tracking LLM Overthinking via Reasoning Dynamics Analysis
by: Wei, Zihao, et al.
Published: (2025)

Don't Overthink it. Preferring Shorter Thinking Chains for Improved LLM Reasoning
by: Hassid, Michael, et al.
Published: (2025)

Reasoning or Overthinking: Evaluating Large Language Models on Financial Sentiment Analysis
by: Vamvourellis, Dimitris, et al.
Published: (2025)

DP-Fusion: Token-Level Differentially Private Inference for Large Language Models
by: Thareja, Rushil, et al.
Published: (2025)

Adaptive Conformal Prediction for Improving Factuality of Generations by Large Language Models
by: Rubashevskii, Aleksandr, et al.
Published: (2026)

I Could've Asked That: Reformulating Unanswerable Questions
by: Zhao, Wenting, et al.
Published: (2024)

Precedent-Informed Reasoning: Mitigating Overthinking in Large Reasoning Models via Test-Time Precedent Learning
by: Wang, Qianyue, et al.
Published: (2026)

Can LLMs Automate Fact-Checking Article Writing?
by: Sahnan, Dhruv, et al.
Published: (2025)

Your Models Have Thought Enough: Training Large Reasoning Models to Stop Overthinking
by: Han, Jinyi, et al.
Published: (2025)

A Survey of Confidence Estimation and Calibration in Large Language Models
by: Geng, Jiahui, et al.
Published: (2023)

Mitigating Overthinking in Large Reasoning Models via Manifold Steering
by: Huang, Yao, et al.
Published: (2025)

EngTrace: A Symbolic Benchmark for Verifiable Process Supervision of Engineering Reasoning
by: Gull, Ayesha, et al.
Published: (2025)

OpenFactCheck: A Unified Framework for Factuality Evaluation of LLMs
by: Iqbal, Hasan, et al.
Published: (2024)

Hop, Skip, and Overthink: Diagnosing Why Reasoning Models Fumble during Multi-Hop Analysis
by: Yadav, Anushka, et al.
Published: (2025)

CLIPer: Tailoring Diverse User Preference via Classifier-Guided Inference-Time Personalization
by: Su, Jinyan, et al.
Published: (2026)

The Thinking Spectrum: An Empirical Study of Tunable Reasoning in LLMs through Model Merging
by: Lan, Xiaochong, et al.
Published: (2025)

Don't "Overthink" Passage Reranking: Is Reasoning Truly Necessary?
by: Jedidi, Nour, et al.
Published: (2025)

Overthinking Reduction with Decoupled Rewards and Curriculum Data Scheduling
by: Jiang, Shuyang, et al.
Published: (2025)

MGM: Global Understanding of Audience Overlap Graphs for Predicting the Factuality and the Bias of News Media
by: Manzoor, Muhammad Arslan, et al.
Published: (2024)