HOME ｜国立情報学研究所大規模言語モデル研究開発センター

By bringing together the research capabilities of industry, academia, and government, we aim to build an academic research hub. We will develop an environment conducive to the cultivation of research and development capabilities related to generative AI models, ensure transparency through elucidation of the learning principles of generative AI models, and carry out R&D that contributes to the advancement of generative AI models.
Through these initiatives, we contribute to the evolution of AI and, ultimately, the creation of innovative breakthroughs over the long term.

Topics

*Scrollable horizontally.

All

Achievements

LLM-jp

Seminars

News

Events

2024.12.24 Achievements

A Fully Open Large Language Model with Approximately 172 Billion Parameters (GPT-3 Level): “llm-jp-3-172b-instruct3” Now Publicly Available – Achieving Performance Beyond GPT-3.5 –

2025.01.17 LLM-jp

Announcement of the 16th LLM-jp Meeting

2025.01.17 LLM-jp

The 15th LLM-jp Meeting

2024.12.27 LLM-jp

Announcement of the 15th LLM-jp Meeting

2024.11.26 LLM-jp

The 14th LLM-jp Meeting

2024.12.24 Achievements

A Fully Open Large Language Model with Approximately 172 Billion Parameters (GPT-3 Level): “llm-jp-3-172b-instruct3” Now Publicly Available – Achieving Performance Beyond GPT-3.5 –

2024.11.20 Achievements

Full-Scratch Learning of a Large Language Model with Approximately 172 billion Parameters (GPT-3 Level) and Preview Release- The World’s Largest Fully Open Model, Including Training Data –

2025.01.17 LLM-jp

GENIAC matching event (2nd)

2024.07.16 Events

NII/LLMC Invited Talk “Cognitive Aspects of Large Language Models” by Prof. Marc Cavazza

In this consortium, researchers in natural language processing and computer systems gather regularly to share information on the research and development of large language models.

Please see the following schedule for this year's study sessions.

5/28

火

14:30-18:00

6/25

火

14:30-18:00

7/30

火

14:30-18:00

8/27

火

14:30-18:00

10/29

火

14:30-18:00

11/26

火

14:30-18:00

1/14

火

14:30-18:00

2/25

火

14:30-18:00

3/25

火

14:30-18:00

Topics

A Fully Open Large Language Model with Approximately 172 Billion Parameters (GPT-3 Level): “llm-jp-3-172b-instruct3” Now Publicly Available – Achieving Performance Beyond GPT-3.5 –

Announcement of the 16th LLM-jp Meeting

The 15th LLM-jp Meeting

Announcement of the 15th LLM-jp Meeting

The 14th LLM-jp Meeting

A Fully Open Large Language Model with Approximately 172 Billion Parameters (GPT-3 Level): “llm-jp-3-172b-instruct3” Now Publicly Available – Achieving Performance Beyond GPT-3.5 –

Announcement of LLM-jp-3 VILA 14B

Release of LLM-jp-3 172B beta2

Release of LLM-jp-3 172B alpha1 and alpha2

Full-Scratch Learning of a Large Language Model with Approximately 172 billion Parameters (GPT-3 Level) and Preview Release- The World’s Largest Fully Open Model, Including Training Data –

Announcement of the 16th LLM-jp Meeting

The 15th LLM-jp Meeting

Announcement of the 15th LLM-jp Meeting

The 14th LLM-jp Meeting

Announcement of the 14th LLM-jp Meeting

LLM-jp won the 19th AAMT Nagao Award

NII/LLMC Invited Talk “Foundation Models for Medicine (and Education) by Prof. Dr. Wolfgang Nejdl

LLM Symposium 2024: What Should Large Language Models Be Like?

LLM-jp Technical Discussion “Technical Discussion with Jonathan Frankle”

GENIAC matching event (2nd)

NII/LLMC Invited Talk “Cognitive Aspects of Large Language Models” by Prof. Marc Cavazza

LLM-jp

In this consortium, researchers in natural language processing and computer systems gather regularly to share information on the research and development of large language models.

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.