EXAONE 3.5: Series of Large Language Models for Real-world Use Cases
LG AI Research, Soyoung An, Kyunghoon Bae, Eunbi Choi, Kibong Choi, Stanley Jungkyu Choi, Seokhee Hong, Junwon Hwang, Hyojin Jeon, Gerrard Jeongwon Jo, Hyunjik Jo, Jiyeon Jung, Yountae Jung, Hyosang Kim, Joonkee Kim, Seonghwan Kim, Soyeon Kim, Sunkyoung Kim, Yireun Kim, Yongil Kim, Youchul Kim, Edward Hwayoung Lee
2024-12-09
Summary
This paper talks about EXAONE 3.5, a series of advanced language models developed by LG AI Research that are designed to understand and process information effectively in real-world situations.
What's the problem?
While many language models exist, they often struggle with following instructions accurately and understanding long texts. This can limit their usefulness in practical applications where clear communication and comprehension are essential.
What's the solution?
The authors introduced EXAONE 3.5, which comes in three different sizes (2.4 billion, 7.8 billion, and 32 billion parameters) to cater to various needs, from lightweight applications to more powerful tasks. These models have been fine-tuned to follow instructions better and understand longer contexts, achieving top scores in multiple benchmarks that test their performance in real-world scenarios.
Why it matters?
This research is important because it enhances the capabilities of open-source language models, making them more accessible for researchers and developers. By improving how these models work with real-world tasks, EXAONE 3.5 can contribute to advancements in areas like customer service, education, and content creation, ultimately making technology more effective and user-friendly.
Abstract
This technical report introduces the EXAONE 3.5 instruction-tuned language models, developed and released by LG AI Research. The EXAONE 3.5 language models are offered in three configurations: 32B, 7.8B, and 2.4B. These models feature several standout capabilities: 1) exceptional instruction following capabilities in real-world scenarios, achieving the highest scores across seven benchmarks, 2) outstanding long-context comprehension, attaining the top performance in four benchmarks, and 3) competitive results compared to state-of-the-art open models of similar sizes across nine general benchmarks. The EXAONE 3.5 language models are open to anyone for research purposes and can be downloaded from https://huggingface.co/LGAI-EXAONE. For commercial use, please reach out to the official contact point of LG AI Research: contact_us@lgresearch.ai.