Defragmenting Language Models: An Interpretability-based Approach for Vocabulary Expansion
cs.CL
Submitted: 2026-04-17
Updated: 2026-11-01
Code: https://github.com/utahnlp/fragmend
Project page: https://mayhewsw.github.io/2016/02/18/how-many-languages-per-script
Terminology
Sources
- One Tokenizer To Rule Them All: Emergent Language Plasticity via Multilingual Tokenizers
- Bridging the Gap for Tokenizer-Free Language Models
- Reducing Tokenization Premiums for Low-Resource Languages
- Efficient and Effective Text Encoding for Chinese LLaMA and Alpaca
- HyperNetworks
- Model Internal Sleuthing: Finding Lexical Identity and Inflectional Features in Modern Language Models
- Tiny Aya: Bridging Scale and Multilingual Depth
- Qwen3 Technical Report
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data