4 papers
Instability in Downstream Task Performance During LLM Pretraining
Yuto Nishida, Masaru Isonuma, Yusuke Oda
When training large language models (LLMs), it is common practice to track downstream task performance throughout the training process and select the checkpoint with the highest va…
Massive Supervised Fine-tuning Experiments Reveal How Data, Layer, and Training Factors Shape LLM Alignment Quality
Yuto Harada, Yusuke Yamauchi, Yusuke Oda +3
Supervised fine-tuning (SFT) is a critical step in aligning large language models (LLMs) with human instructions and values, yet many aspects of SFT remain poorly understood. We tr…
llm-jp-modernbert: A ModernBERT Model Trained on a Large-Scale Japanese Corpus with Long Context Length
Issa Sugiura, Kouta Nakayama, Yusuke Oda
Encoder-only transformer models like BERT are widely adopted as a pre-trained backbone for tasks like sentence classification and retrieval. However, pretraining of encoder models…
Drop-Upcycling: Training Sparse Mixture of Experts with Partial Re-initialization
Taishi Nakamura, Takuya Akiba, Kazuki Fujii +3
The Mixture of Experts (MoE) architecture reduces the training and inference cost significantly compared to a dense model of equivalent capacity. Upcycling is an approach that init…