L

Llama 3 Open Ko 8B Instruct Preview

Developed by beomi
A Korean language model based on continued pre-training of Llama-3-8B, trained on 60GB+ deduplicated publicly available text, supporting Korean and English.
Downloads 6,014
Release Time : 4/23/2024

Model Overview

This is a Korean language model based on the Llama-3-8B architecture, further pre-trained and instruction-tuned for Korean text generation tasks.

Model Features

Korean optimization
Specially optimized for Korean language processing using the new Llama-3 tokenizer for Korean text.
Instruction following
Instruction-tuned using the Chat Vector paper method for better adherence to user instructions.
Public data training
Trained exclusively on 60GB+ deduplicated publicly available text to ensure transparent data sources.

Model Capabilities

Korean text generation
English text generation
Code generation
Q&A systems
Knowledge explanation

Use Cases

Education
Korean language teaching aid
Helps students understand Korean grammar and vocabulary
Provides accurate Korean explanations and examples
Programming
Code generation
Generates Python code based on Korean descriptions
Demonstrates multiple implementations of Fibonacci sequence
History
Historical figure explanations
Detailed explanations of Korean historical figures and events
Provides comprehensive introduction about Admiral Yi Sun-sin
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase