2607.15232v1
CC BY 4.0
MIT
預訓練 LLM 的原地詞元化器擴充
In-Place Tokenizer Expansion for Pre-trained LLMs
Jimmy T. H. Smith, Tarek Dakhran, Alberto Cabrera, Simon S. Lee, Paul Pak, Aditya Tadimeti, Tim Seyd...
LLMTokenizer詞彙擴充多語言持續預訓練On-deviceMoE
Liang-Hsun Huang 許願