Cover Image

How Language Directions Align with Token Geometry in Multilingual LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Kim, JaeSeong, Lee, Suan
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Smoothie-Qwen: Post-Hoc Smoothing to Reduce Language Bias in Multilingual LLMs
by: Ji, SeungWon, et al.
Published: (2025)

How does a Language-Specific Tokenizer affect LLMs?
by: Seo, Jean, et al.
Published: (2025)

Accelerating Multilingual Language Model for Excessively Tokenized Languages
by: Hong, Jimin, et al.
Published: (2024)

Controlling Language Confusion in Multilingual LLMs
by: Lee, Nahyun, et al.
Published: (2025)

AlignX: Advancing Multilingual Large Language Models with Multilingual Representation Alignment
by: Bu, Mengyu, et al.
Published: (2025)

LatinX: Aligning a Multilingual TTS Model with Direct Preference Optimization
by: Chary, Luis Felipe, et al.
Published: (2025)

Improving Multilingual Language Models by Aligning Representations through Steering
by: Mahmoud, Omar, et al.
Published: (2025)

Beyond Early-Token Bias: Model-Specific and Language-Specific Position Effects in Multilingual LLMs
by: Menschikov, Mikhail, et al.
Published: (2025)

What is Your Favorite Gender, MLM? Gender Bias Evaluation in Multilingual Masked Language Models
by: Yu, Jeongrok, et al.
Published: (2024)

One Tokenizer To Rule Them All: Emergent Language Plasticity via Multilingual Tokenizers
by: Abagyan, Diana, et al.
Published: (2025)

WoLF: Wide-scope Large Language Model Framework for CXR Understanding
by: Kang, Seil, et al.
Published: (2024)

TASE: Token Awareness and Structured Evaluation for Multilingual Language Models
by: Zhao, Chenzhuo, et al.
Published: (2025)

Multilingual Tokenization through the Lens of Indian Languages: Challenges and Insights
by: Karthika, N J, et al.
Published: (2025)

How Does Quantization Affect Multilingual LLMs?
by: Marchisio, Kelly, et al.
Published: (2024)

CM-Align: Consistency-based Multilingual Alignment for Large Language Models
by: Zhang, Xue, et al.
Published: (2025)

Improving Conversational Abilities of Quantized Large Language Models via Direct Preference Alignment
by: Lee, Janghwan, et al.
Published: (2024)

Beyond Literal Token Overlap: Token Alignability for Multilinguality
by: Hämmerl, Katharina, et al.
Published: (2025)

TASTE-Streaming: Towards Streamable Text-Aligned Speech Tokenization and Embedding for Spoken Language Modeling
by: Tseng, Liang-Hsuan, et al.
Published: (2026)

Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs
by: Weeber, Franziska, et al.
Published: (2025)

Optimizing Korean-Centric LLMs via Token Pruning
by: Kim, Hoyeol, et al.
Published: (2026)

Aligning Language Models to Explicitly Handle Ambiguity
by: Kim, Hyuhng Joon, et al.
Published: (2024)

TASTE: Text-Aligned Speech Tokenization and Embedding for Spoken Language Modeling
by: Tseng, Liang-Hsuan, et al.
Published: (2025)

From Unaligned to Aligned: Scaling Multilingual LLMs with Multi-Way Parallel Corpora
by: Shen, Yingli, et al.
Published: (2025)

How Does Alignment Enhance LLMs' Multilingual Capabilities? A Language Neurons Perspective
by: Zhang, Shimao, et al.
Published: (2025)

Less Data Less Tokens: Multilingual Unification Learning for Efficient Test-Time Reasoning in LLMs
by: Chen, Kang, et al.
Published: (2025)

AlignFreeze: Navigating the Impact of Realignment on the Layers of Multilingual Models Across Diverse Languages
by: Bakos, Steve, et al.
Published: (2025)

CORAL: Adaptive Retrieval Loop for Culturally-Aligned Multilingual RAG
by: Lee, Nayeon, et al.
Published: (2026)

Turning English-centric LLMs Into Polyglots: How Much Multilinguality Is Needed?
by: Kew, Tannon, et al.
Published: (2023)

Do LLMs Truly Benefit from Longer Context in Automatic Post-Editing?
by: Kim, Ahrii, et al.
Published: (2026)

Pragmatic inference of scalar implicature by LLMs
by: Cho, Ye-eun, et al.
Published: (2024)

I Am Aligned, But With Whom? MENA Values Benchmark for Evaluating Cultural Alignment and Multilingual Bias in LLMs
by: Zahraei, Pardis Sadat, et al.
Published: (2025)

KiC: Keyword-inspired Cascade for Cost-Efficient Text Generation with LLMs
by: Kim, Woo-Chan, et al.
Published: (2025)

Pruning Multilingual Large Language Models for Multilingual Inference
by: Kim, Hwichan, et al.
Published: (2024)

A Shared Geometry of Difficulty in Multilingual Language Models
by: Civelli, Stefano, et al.
Published: (2026)

The Token Tax: Systematic Bias in Multilingual Tokenization
by: Lundin, Jessica M., et al.
Published: (2025)

Language Bias under Conflicting Information in Multilingual LLMs
by: Östling, Robert, et al.
Published: (2026)

Investigating the Influence of Language on Sycophantic Behavior of Multilingual LLMs
by: Aldahlawi, Bayan Abdullah, et al.
Published: (2026)

Refusal Direction is Universal Across Safety-Aligned Languages
by: Wang, Xinpeng, et al.
Published: (2025)

Registering Source Tokens to Target Language Spaces in Multilingual Neural Machine Translation
by: Qu, Zhi, et al.
Published: (2025)

MAGNET: Improving the Multilingual Fairness of Language Models with Adaptive Gradient-Based Tokenization
by: Ahia, Orevaoghene, et al.
Published: (2024)