Japanese Large Language Model
Research on Japanese Large Language Models (JLLMs) focuses on improving their performance across various tasks, including biomedical applications, speech recognition, and question answering, often using instruction tuning and continual pre-training methods. Current efforts concentrate on developing robust benchmarks for fair evaluation, addressing biases, and enhancing cross-lingual capabilities by leveraging existing English resources while mitigating potential cultural misalignment. These advancements are significant for both advancing natural language processing research and enabling practical applications in diverse fields like healthcare and finance within the Japanese language context.
Papers
Continual Pre-Training for Cross-Lingual LLM Adaptation: Enhancing Japanese Language Capabilities
Kazuki Fujii, Taishi Nakamura, Mengsay Loem, Hiroki Iida, Masanari Ohi, Kakeru Hattori, Hirai Shota, Sakae Mizuki, Rio Yokota, Naoaki Okazaki
Building a Large Japanese Web Corpus for Large Language Models
Naoaki Okazaki, Kakeru Hattori, Hirai Shota, Hiroki Iida, Masanari Ohi, Kazuki Fujii, Taishi Nakamura, Mengsay Loem, Rio Yokota, Sakae Mizuki