EXAONE 3.0 7.8B Instruction Tuned Language Model
LG AI Research, Soyoung An, Kyunghoon Bae, Eunbi Choi, Stanley Jungkyu Choi, Yemuk Choi, Seokhee Hong, Yeonjung Hong, Junwon Hwang, Hyojin Jeon, Gerrard Jeongwon Jo, Hyunjik Jo, Jiyeon Jung, Yountae Jung, Euisoon Kim, Hyosang Kim, Joonkee Kim, Seonghwan Kim, Soyeon Kim, Sunkyoung Kim, Yireun Kim, Youchul Kim
2024-08-08
Summary
This paper introduces EXAONE 3.0, an instruction-tuned language model developed by LG AI Research, which is designed to perform well in understanding and generating text, especially in both English and Korean.
What's the problem?
Large language models (LLMs) are often either closed-source or not optimized for specific tasks, leading to a performance gap between powerful models and those available for public use. This can make it difficult for researchers and developers to access high-quality tools that can handle complex language tasks effectively.
What's the solution?
The authors released the EXAONE 3.0 model, specifically the 7.8 billion parameter version, as an open-source tool to promote research and innovation. This model has been fine-tuned to follow instructions better and has shown strong performance in various benchmarks, particularly excelling in Korean language tasks while also being effective in English. The model was trained on a massive dataset of 8 trillion tokens, allowing it to learn from a wide range of topics and styles.
Why it matters?
EXAONE 3.0 is important because it provides a high-performance language model that is accessible to the public, fostering collaboration and innovation in AI research. By focusing on bilingual capabilities, particularly in Korean, this model can help advance natural language processing in regions where such tools are needed most, making it easier for developers to create applications that require understanding and generating text.
Abstract
We introduce EXAONE 3.0 instruction-tuned language model, the first open model in the family of Large Language Models (LLMs) developed by LG AI Research. Among different model sizes, we publicly release the 7.8B instruction-tuned model to promote open research and innovations. Through extensive evaluations across a wide range of public and in-house benchmarks, EXAONE 3.0 demonstrates highly competitive real-world performance with instruction-following capability against other state-of-the-art open models of similar size. Our comparative analysis shows that EXAONE 3.0 excels particularly in Korean, while achieving compelling performance across general tasks and complex reasoning. With its strong real-world effectiveness and bilingual proficiency, we hope that EXAONE keeps contributing to advancements in Expert AI. Our EXAONE 3.0 instruction-tuned model is available at https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct