Image de couverture de livre

Lost in Multilinguality: Dissecting Cross-lingual Factual Inconsistency in Transformer Language Models

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Wang, Mingyang, Adel, Heike, Lange, Lukas, Liu, Yihong, Nie, Ercong, Strötgen, Jannik, Schütze, Hinrich
Format:	Preprint
Publié:	2025
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

NLNDE at SemEval-2023 Task 12: Adaptive Pretraining and Source Language Selection for Low-Resource Multilingual Sentiment Analysis
par: Wang, Mingyang, et autres
Publié: (2023)

Rehearsal-Free Modular and Compositional Continual Learning for Language Models
par: Wang, Mingyang, et autres
Publié: (2024)

Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization
par: Wang, Mingyang, et autres
Publié: (2024)

Language Mixing in Reasoning Language Models: Patterns, Impact, and Internal Causes
par: Wang, Mingyang, et autres
Publié: (2025)

Learn it or Leave it: Module Composition and Pruning for Continual Learning
par: Wang, Mingyang, et autres
Publié: (2024)

Bring Your Own Knowledge: A Survey of Methods for LLM Knowledge Expansion
par: Wang, Mingyang, et autres
Publié: (2025)

Tracing Multilingual Factual Knowledge Acquisition in Pretraining
par: Liu, Yihong, et autres
Publié: (2025)

Evaluating Contextually Mediated Factual Recall in Multilingual Large Language Models
par: Liu, Yihong, et autres
Publié: (2026)

Mechanistic Understanding and Mitigation of Language Confusion in English-Centric Large Language Models
par: Nie, Ercong, et autres
Publié: (2025)

LangSAMP: Language-Script Aware Multilingual Pretraining
par: Liu, Yihong, et autres
Publié: (2024)

OFA: A Framework of Initializing Unseen Subword Embeddings for Efficient Large-scale Multilingual Continued Pretraining
par: Liu, Yihong, et autres
Publié: (2023)

Breaking the Script Barrier in Multilingual Pre-Trained Language Models with Transliteration-Based Post-Training Alignment
par: Xhelili, Orgest, et autres
Publié: (2024)

BMIKE-53: Investigating Cross-Lingual Knowledge Editing with In-Context Learning
par: Nie, Ercong, et autres
Publié: (2024)

Decomposed Prompting: Probing Multilingual Linguistic Structure Knowledge in Large Language Models
par: Nie, Ercong, et autres
Publié: (2024)

DiM\textsuperscript{3}: Bridging Multilingual and Multimodal Models via Direction- and Magnitude-Aware Merging
par: Wang, Zijing, et autres
Publié: (2026)

TransliCo: A Contrastive Learning Framework to Address the Script Barrier in Multilingual Pretrained Language Models
par: Liu, Yihong, et autres
Publié: (2024)

TransMI: A Framework to Create Strong Baselines from Multilingual Pretrained Language Models for Transliterated Data
par: Liu, Yihong, et autres
Publié: (2024)

GNNavi: Navigating the Information Flow in Large Language Models by Graph Neural Network
par: Yuan, Shuzhou, et autres
Publié: (2024)

Why Better Cross-Lingual Alignment Fails for Better Cross-Lingual Transfer: Case of Encoders
par: Veitsman, Yana, et autres
Publié: (2026)

On the Entity-Level Alignment in Crosslingual Consistency
par: Liu, Yihong, et autres
Publié: (2025)

Refusal Direction is Universal Across Safety-Aligned Languages
par: Wang, Xinpeng, et autres
Publié: (2025)

Large Reasoning Models Are (Not Yet) Multilingual Latent Reasoners
par: Liu, Yihong, et autres
Publié: (2026)

Hateful Person or Hateful Model? Investigating the Role of Personas in Hate Speech Detection by Large Language Models
par: Yuan, Shuzhou, et autres
Publié: (2025)

Large Language Models as Neurolinguistic Subjects: Discrepancy between Performance and Competence
par: He, Linyang, et autres
Publié: (2024)

Evaluating Robustness of Large Language Models Against Multilingual Typographical Errors
par: Zhao, Raoyuan, et autres
Publié: (2025)

A Comprehensive Evaluation of Multilingual Chain-of-Thought Reasoning: Performance, Consistency, and Faithfulness Across Languages
par: Zhao, Raoyuan, et autres
Publié: (2025)

Discourse-Aware In-Context Learning for Temporal Expression Normalization
par: Gautam, Akash Kumar, et autres
Publié: (2024)

Crosslingual On-Policy Self-Distillation for Multilingual Reasoning
par: Liu, Yihong, et autres
Publié: (2026)

GKnow: Measuring the Entanglement of Gender Bias and Factual Gender
par: Veloso, Leonor, et autres
Publié: (2026)

Your Pretrained Model Tells the Difficulty Itself: A Self-Adaptive Curriculum Learning Paradigm for Natural Language Understanding
par: Feng, Qi, et autres
Publié: (2025)

How Programming Concepts and Neurons Are Shared in Code Language Models
par: Kargaran, Amir Hossein, et autres
Publié: (2025)

ToPro: Token-Level Prompt Decomposition for Cross-Lingual Sequence Labeling Tasks
par: Ma, Bolei, et autres
Publié: (2024)

HYPEROFA: Expanding LLM Vocabulary to New Languages via Hypernetwork-Based Embedding Initialization
par: Özeren, Enes, et autres
Publié: (2025)

SAD: A Large-Scale Strategic Argumentative Dialogue Dataset
par: Liu, Yongkang, et autres
Publié: (2026)

GLUScope: A Tool for Analyzing GLU Neurons in Transformer Language Models
par: Gerstner, Sebastian, et autres
Publié: (2026)

A Recipe of Parallel Corpora Exploitation for Multilingual Large Language Models
par: Lin, Peiqin, et autres
Publié: (2024)

Relational Linearity is a Predictor of Hallucinations
par: Lu, Yuetian, et autres
Publié: (2026)

Evaluating Knowledge-based Cross-lingual Inconsistency in Large Language Models
par: Xing, Xiaolin, et autres
Publié: (2024)

LLM in the Loop: Creating the ParaDeHate Dataset for Hate Speech Detoxification
par: Yuan, Shuzhou, et autres
Publié: (2025)

The Anatomy of an Edit: Mechanism-Guided Activation Steering for Knowledge Editing
par: Cao, Yuan, et autres
Publié: (2026)