Imagen de Portada

Probing the Geometry of Truth: Consistency and Generalization of Truth Directions in LLMs Across Logical Transformations and Question Answering Tasks

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Bao, Yuntai, Zhang, Xuhong, Du, Tianyu, Zhao, Xinkui, Feng, Zhengwen, Peng, Hao, Yin, Jianwei
Formato:	Preprint
Publicado:	2025
Materias:	Computation and Language
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Scalable Multi-Stage Influence Function for Large Language Models via Eigenvalue-Corrected Kronecker-Factored Parameterization
por: Bao, Yuntai, et al.
Publicado: (2025)

HijackRAG: Hijacking Attacks against Retrieval-Augmented Large Language Models
por: Zhang, Yucheng, et al.
Publicado: (2024)

The Geometries of Truth Are Orthogonal Across Tasks
por: Azizian, Waiss, et al.
Publicado: (2025)

Faithful Bi-Directional Model Steering via Distribution Matching and Distributed Interchange Interventions
por: Bao, Yuntai, et al.
Publicado: (2026)

Testing the Limits of Truth Directions in LLMs
por: Poulis, Angelos, et al.
Publicado: (2026)

Pruning Weights but Not Truth: Safeguarding Truthfulness While Pruning LLMs
por: Fu, Yao, et al.
Publicado: (2025)

How Context Shapes Truth: Geometric Transformations of Statement-level Truth Representations in LLMs
por: Adarsh, Shivam, et al.
Publicado: (2026)

Logical Form and Truth-Conditions
por: Andrea IACONA
Publicado: (2013)

Truth Claims Across Media
Publicado: (2024)

Autonomous Evaluation of LLMs for Truth Maintenance and Reasoning Tasks
por: Karia, Rushang, et al.
Publicado: (2024)

TruthRL: Incentivizing Truthful LLMs via Reinforcement Learning
por: Wei, Zhepei, et al.
Publicado: (2025)

Safeguarding the Truth of High-Value Price Oracle Task: A Dynamically Adjusted Truth Discovery Method
por: Xian, Youquan, et al.
Publicado: (2024)

The Muses of Truth and Transformation
por: Chinen, Allan B.
Publicado: (2024)

Uhura: A Benchmark for Evaluating Scientific Question Answering and Truthfulness in Low-Resource African Languages
por: Bayes, Edward, et al.
Publicado: (2024)

Towards Steering without Sacrifice: Principled Training of Steering Vectors for Prompt-only Interventions
por: Bao, Yuntai, et al.
Publicado: (2026)

ERA-CoT: Improving Chain-of-Thought through Entity Relationship Analysis
por: Liu, Yanming, et al.
Publicado: (2024)

Compressing LLMs: The Truth is Rarely Pure and Never Simple
por: Jaiswal, Ajay, et al.
Publicado: (2023)

CoreGuard: Safeguarding Foundational Capabilities of LLMs Against Model Stealing in Edge Deployment
por: Li, Qinfeng, et al.
Publicado: (2024)

EdgeJury: Cross-Reviewed Small-Model Ensembles for Truthful Question Answering on Serverless Edge Inference
por: Kumar, Aayush
Publicado: (2025)

TransLinkGuard: Safeguarding Transformer Models Against Model Stealing in Edge Deployment
por: Li, Qinfeng, et al.
Publicado: (2024)

Tool-Planner: Task Planning with Clusters across Multiple Tools
por: Liu, Yanming, et al.
Publicado: (2024)

On the Universal Truthfulness Hyperplane Inside LLMs
por: Liu, Junteng, et al.
Publicado: (2024)

Walking the Schrödinger Bridge: A Direct Trajectory for Text-to-3D Generation
por: Li, Ziying, et al.
Publicado: (2025)

TruthFlow: Truthful LLM Generation via Representation Flow Correction
por: Wang, Hanyu, et al.
Publicado: (2025)

TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space
por: Zhang, Shaolei, et al.
Publicado: (2024)

Love the Truth, the Whole Truth, and the Truth about Everything. An Interview with Josef Seifert
por: Rodrigo Guerra López
Publicado: (2014)

MIDAS: Modeling Ground-Truth Distributions with Dark Knowledge for Domain Generalized Stereo Matching
por: Xu, Peng, et al.
Publicado: (2025)

Queries With Exact Truth Values in Paraconsistent Description Logics
por: Bienvenu, Meghyn, et al.
Publicado: (2024)

Accurate Table Question Answering with Accessible LLMs
por: Jiang, Yangfan, et al.
Publicado: (2026)

Truth
por: Cubitt, Sean
Publicado: (2024)

Ground Truth Generation for Multilingual Historical NLP using LLMs
por: Gladstone, Clovis, et al.
Publicado: (2025)

The Truth, the Whole Truth, and Nothing but the Truth: Automatic Visualization Evaluation from Reconstruction Quality
por: Bujack, Roxana, et al.
Publicado: (2026)

Truth-Aware Decoding: A Program-Logic Approach to Factual Language Generation
por: Alpay, Faruk, et al.
Publicado: (2025)

Compositional Consistency-Guided Decoding for Three-Way Logical Question Answering
por: Huang, Tianyi, et al.
Publicado: (2026)

Truthful Aggregation of LLMs with an Application to Online Advertising
por: Soumalias, Ermis, et al.
Publicado: (2024)

Truth is Universal: Robust Detection of Lies in LLMs
por: Bürger, Lennart, et al.
Publicado: (2024)

RA-ISF: Learning to Answer and Understand from Retrieval Augmentation via Iterative Self-Feedback
por: Liu, Yanming, et al.
Publicado: (2024)

Neural Quantum States in Variational Monte Carlo Method: A Brief Summary
por: Song, Yuntai
Publicado: (2024)

Truth Knows No Language: Evaluating Truthfulness Beyond English
por: Figueras, Blanca Calvo, et al.
Publicado: (2025)

TruthStance: An Annotated Dataset of Conversations on Truth Social
por: Ameen, Fathima, et al.
Publicado: (2026)