Swallow LLM

Swallow LLM

Research and development of large language models conducted by the members mainly in Okazaki Laboratory and Yokota Laboratory at Tokyo Institute of Technology

View on HuggingFace

Swallow LLM

Project

Llama 3 Swallow

Llama 3 Swallow

LLMs that enhance Japanese capability of Llama 3 (8B, 70B)

Swallow on Mistral

Swallow on Mistral

LLMs that enhance Japanese capability of Mistral 7B and Mixtral 8x7B

Swallow

Swallow

LLMs that enhance Japanese capability of Llama 2 (7B, 13B, 70B)

Member

Naoaki Okazaki

Professor, TokyoTech

Project leader, team leader of pre-training corpus, Web developer

Rio Yokota

Professor, TokyoTech

Team leader of LLM training

Sakae Mizuki

Researcher, AIST / TokyoTech

Team leader of instruction tuning, team leader of LLM evaluation

Kazuki Fujii

Master student, TokyoTech

LLM training, instruction tuning

Taishi Nakamura

Master student, TokyoTech

LLM training, instruction tuning, and evaluation

Youmi Ma

PhD student, TokyoTech

LLM evaluation

Koki Maeda

PhD student, TokyoTech

LLM evaluation

Kakeru Hattori

Master student, TokyoTech

Development of pre-training corpora, LLM evaluation

Masanari Ohi

Master student, TokyoTech

LLM evaluation

Takumi Okamoto

Master student, TokyoTech

Instruction tuning

Ishida Shigeki

Master student, TokyoTech

LLM evaluation

Taihei Shiotani

Master student, TokyoTech

LLM evaluation

Koshiro Saito

Undergraduate student, TokyoTech

LLM evaluation, exploration of tokenization

Hiroya Takamura

Research team leader, AIRC, AIST

Project manager

Past Member

Mengsay Loem

Master of TokyoTech

Expert of LLM evaluation

Shota Hirai

Master of TokyoTech

Expert of development of pre-training corpora

Taiki Iida

PhD of TokyoTech

Expoert of tokenization for LLMs