Image de couverture de livre

Tomato, Tomahto, Tomate: Do Multilingual Language Models Understand Based on Subword-Level Semantic Concepts?

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhang, Crystina, Lu, Jing, Tran, Vinh Q., Schuster, Tal, Metzler, Donald, Lin, Jimmy
Format:	Preprint
Publié:	2024
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

The Role of Mixed-Language Documents for Multilingual Large Language Model Pretraining
par: Shao, Jiandong, et autres
Publié: (2026)

Hard Negatives, Hard Lessons: Revisiting Training Data Quality for Robust Information Retrieval with LLMs
par: Thakur, Nandan, et autres
Publié: (2025)

Understanding Subword Compositionality of Large Language Models
par: Peng, Qiwei, et autres
Publié: (2025)

StochasTok: Improving Fine-Grained Subword Understanding in LLMs
par: Sims, Anya, et autres
Publié: (2025)

A Systematic Analysis of Subwords and Cross-Lingual Transfer in Multilingual Translation
par: Meyer, Francois, et autres
Publié: (2024)

CSF: Contrastive Semantic Features for Direct Multilingual Sign Language Generation
par: Bao, Tran Sy
Publié: (2026)

LACONIC: Dense-Level Effectiveness for Scalable Sparse Retrieval via a Two-Phase Training Curriculum
par: Xu, Zhichao, et autres
Publié: (2026)

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

Impact of Preference Noise on the Alignment Performance of Generative Language Models
par: Gao, Yang, et autres
Publié: (2024)

Rankers, Judges, and Assistants: Towards Understanding the Interplay of LLMs in Information Retrieval Evaluation
par: Balog, Krisztian, et autres
Publié: (2025)

Can Pretrained Language Models Derive Correct Semantics from Corrupt Subwords under Noise?
par: Li, Xinzhe, et autres
Publié: (2023)

Drawing Conclusions from Draws: Rethinking Preference Semantics in Arena-Style LLM Evaluation
par: Tang, Raphael, et autres
Publié: (2025)

SEMQA: Semi-Extractive Multi-Source Question Answering
par: Schuster, Tal, et autres
Publié: (2023)

Do Language Models' Words Refer?
par: Mandelkern, Matthew, et autres
Publié: (2023)

Fractal Patterns May Illuminate the Success of Next-Token Prediction
par: Alabdulmohsin, Ibrahim, et autres
Publié: (2024)

The Learning Dynamics of Subword Segmentation for Morphologically Diverse Languages
par: Meyer, Francois, et autres
Publié: (2025)

Morphological Typology in BPE Subword Productivity and Language Modeling
par: Parra, Iñigo
Publié: (2024)

Evaluating Subword Tokenization: Alien Subword Composition and OOV Generalization Challenge
par: Batsuren, Khuyagbaatar, et autres
Publié: (2024)

Improving Multilingual Math Reasoning for African Languages
par: Ogundepo, Odunayo, et autres
Publié: (2025)

Distributional Properties of Subword Regularization
par: Cognetta, Marco, et autres
Publié: (2024)

Lexically Grounded Subword Segmentation
par: Libovický, Jindřich, et autres
Publié: (2024)

Tokenization Falling Short: On Subword Robustness in Large Language Models
par: Chai, Yekun, et autres
Publié: (2024)

All Languages Matter: Understanding and Mitigating Language Bias in Multilingual RAG
par: Wang, Dan, et autres
Publié: (2026)

On the Effect of (Near) Duplicate Subwords in Language Modelling
par: Schäfer, Anton, et autres
Publié: (2024)

Entailment Semantics Can Be Extracted from an Ideal Language Model
par: Merrill, William, et autres
Publié: (2022)

ByteSpan: Information-Driven Subword Tokenisation
par: Goriely, Zébulon, et autres
Publié: (2025)

Subword Tokenization Strategies for Kurdish Word Embeddings
par: Salehi, Ali, et autres
Publié: (2025)

Transferring Extreme Subword Style Using Ngram Model-Based Logit Scaling
par: Messner, Craig, et autres
Publié: (2025)

MINERS: Multilingual Language Models as Semantic Retrievers
par: Winata, Genta Indra, et autres
Publié: (2024)

Decoupling the Benefits of Subword Tokenization for Language Model Training via Byte-level Simulation
par: Gigant, Théo, et autres
Publié: (2026)

How Do Multilingual Language Models Remember Facts?
par: Fierro, Constanza, et autres
Publié: (2024)

Do Pre-Trained Language Models Detect and Understand Semantic Underspecification? Ask the DUST!
par: Wildenburg, Frank, et autres
Publié: (2024)

LGSE: Lexically Grounded Subword Embedding Initialization for Low-Resource Language Adaptation
par: Teklehaymanot, Hailay, et autres
Publié: (2026)

Subword models struggle with word learning, but surprisal hides it
par: Bunzeck, Bastian, et autres
Publié: (2025)

Learning Mutually Informed Representations for Characters and Subwords
par: Wang, Yilin, et autres
Publié: (2023)

Leading Whitespaces of Language Models' Subword Vocabulary Pose a Confound for Calculating Word Probabilities
par: Oh, Byung-Doh, et autres
Publié: (2024)

Smaller, Weaker, Yet Better: Training LLM Reasoners via Compute-Optimal Sampling
par: Bansal, Hritik, et autres
Publié: (2024)

Existential Definability over the Subword Ordering
par: Baumann, Pascal, et autres
Publié: (2022)

INCLUDE: Evaluating Multilingual Language Understanding with Regional Knowledge
par: Romanou, Angelika, et autres
Publié: (2024)

Do Multilingual Large Language Models Mitigate Stereotype Bias?
par: Nie, Shangrui, et autres
Publié: (2024)