Image de couverture de livre

What are They Thinking? Delineation, Probing and Tracking of Concepts in LLMs

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Abdelwahab, Mohamed, Collins, Michelle Yu, Chen, Sihan, Zhao, Yi Cheng, Mahmood, Zafarullah, Zhu, Jiading, Ali, Soliman, Rose, Jonathan
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

A Fully Generative Motivational Interviewing Counsellor Chatbot for Moving Smokers Towards the Decision to Quit
par: Mahmood, Zafarullah, et autres
Publié: (2025)

Generation, Distillation and Evaluation of Motivational Interviewing-Style Reflections with a Foundational Language Model
par: Brown, Andrew, et autres
Publié: (2024)

Do We Know What LLMs Don't Know? A Study of Consistency in Knowledge Probing
par: Zhao, Raoyuan, et autres
Publié: (2025)

Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space
par: Zhang, Zhen, et autres
Publié: (2025)

Argument Reconstruction as Supervision for Critical Thinking in LLMs
par: Ryu, Hyun, et autres
Publié: (2026)

Beyond English-Centric LLMs: What Language Do Multilingual Language Models Think in?
par: Zhong, Chengzhi, et autres
Publié: (2024)

Beyond Single Concept Vector: Modeling Concept Subspace in LLMs with Gaussian Distribution
par: Zhao, Haiyan, et autres
Publié: (2024)

Can Thinking Models Think to Detect Hateful Memes?
par: Kmainasi, Mohamed Bayan, et autres
Publié: (2026)

Parallel-Probe: Towards Efficient Parallel Thinking via 2D Probing
par: Zheng, Tong, et autres
Publié: (2026)

Not All Tokens Are What You Need In Thinking
par: Yuan, Hang, et autres
Publié: (2025)

Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?
par: Harada, Yuto, et autres
Publié: (2026)

Thinking to Recall: How Reasoning Unlocks Parametric Knowledge in LLMs
par: Gekhman, Zorik, et autres
Publié: (2026)

Think Together and Work Better: Combining Humans' and LLMs' Think-Aloud Outcomes for Effective Text Evaluation
par: Chu, SeongYeub, et autres
Publié: (2024)

What Makes the Preferred Thinking Direction for LLMs in Multiple-choice Questions?
par: Zhang, Yizhe, et autres
Publié: (2025)

Jinx: Unlimited LLMs for Probing Alignment Failures
par: Zhao, Jiahao, et autres
Publié: (2025)

LLMs Learn Constructions That Humans Do Not Know
par: Dunn, Jonathan, et autres
Publié: (2025)

ICR Probe: Tracking Hidden State Dynamics for Reliable Hallucination Detection in LLMs
par: Zhang, Zhenliang, et autres
Publié: (2025)

Probing Minimalist Phase Structure in LLMs: What Universal Dependencies Cannot Represent
par: Chen, Yuanhao, et autres
Publié: (2026)

Measuring What LLMs Think They Do: SHAP Faithfulness and Deployability on Financial Tabular Classification
par: AlMarri, Saeed, et autres
Publié: (2025)

Domain Knowledge-Enhanced LLMs for Fraud and Concept Drift Detection
par: Şenol, Ali, et autres
Publié: (2025)

Do LLMs Really Think Step-by-step In Implicit Reasoning?
par: Yu, Yijiong
Publié: (2024)

RLVR Training of LLMs Does Not Improve Thinking Ability for General QA: Evaluation Method and a Simple Solution
par: Li, Kaiyuan, et autres
Publié: (2026)

What Am I Missing? Question-Answering as Hidden State Probing
par: Luo, Chu Fei, et autres
Publié: (2026)

LingGym: How Far Are LLMs from Thinking Like Field Linguists?
par: Yang, Changbing, et autres
Publié: (2025)

rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking
par: Guan, Xinyu, et autres
Publié: (2025)

FishBack: Pullback Fisher Geometry for Optimal Activation Steering in Transformers
par: Wang, Sihan, et autres
Publié: (2026)

Do LLMs Know What Luxembourgish Borrows? Probing Lexical Neology in Low-Resource Multilingual Models
par: Hosseini-Kivanani, Nina
Publié: (2026)

Systematic Evaluation of Long-Context LLMs on Financial Concepts
par: Gupta, Lavanya, et autres
Publié: (2024)

How LLMs Might Think
par: Gottlieb, Joseph, et autres
Publié: (2026)

When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs
par: Li, Xiaomin, et autres
Publié: (2025)

Fast, Slow, and Tool-augmented Thinking for LLMs: A Review
par: Jia, Xinda, et autres
Publié: (2025)

On Code-Induced Reasoning in LLMs
par: Waheed, Abdul, et autres
Publié: (2025)

Think More, Hallucinate Less: Mitigating Hallucinations via Dual Process of Fast and Slow Thinking
par: Cheng, Xiaoxue, et autres
Publié: (2025)

Atomic Thinking of LLMs: Decoupling and Exploring Mathematical Reasoning Abilities
par: Kuang, Jiayi, et autres
Publié: (2025)

Multilingual != Multicultural: Evaluating Gaps Between Multilingual Capabilities and Cultural Alignment in LLMs
par: Rystrøm, Jonathan, et autres
Publié: (2025)

Recursive Think-Answer Process for LLMs and VLMs
par: Lee, Byung-Kwan, et autres
Publié: (2026)

ODKE+: Ontology-Guided Open-Domain Knowledge Extraction with LLMs
par: Khorshidi, Samira, et autres
Publié: (2025)

ThinkPatterns-21k: A Systematic Study on the Impact of Thinking Patterns in LLMs
par: Wen, Pengcheng, et autres
Publié: (2025)

Reasoning Models Don't Always Say What They Think
par: Chen, Yanda, et autres
Publié: (2025)

Probe Pruning: Accelerating LLMs through Dynamic Pruning via Model-Probing
par: Le, Qi, et autres
Publié: (2025)