B

Biollama Ko 8B

Developed by iRASC
BioLlama-Ko-8B is a Korean medical large language model based on the Llama-3 architecture, constructed by merging the beomi/Llama-3-Open-Ko-8B and ProbeMedicalYonseiMAILab/medllama3-v20 models, specializing in Korean medical Q&A tasks.
Downloads 1,745
Release Time : 6/5/2024

Model Overview

This model employs the DARE TIES merging method, combining general Korean language understanding with medical expertise, and demonstrates outstanding performance in Korean medical Q&A benchmarks, surpassing similar models.

Model Features

Medical Expertise
Incorporates professional medical knowledge through the merger with the medical specialty model medllama3-v20.
Korean Language Optimization
Based on the Llama-3-Open-Ko-8B model, specifically optimized for the Korean language.
High-Performance Merge
Utilizes the advanced DARE TIES merging method, preserving the strengths of both models.

Model Capabilities

Korean Medical Q&A
Medical Knowledge Reasoning
Medical Multiple-Choice Question Answering

Use Cases

Medical Education
Medical Student Learning Aid
Assists medical students in answering multiple-choice questions and conceptual problems.
Demonstrates excellent performance on the KorMedMCQA benchmark.
Medical Consultation
Basic Medical Q&A
Answers common medical questions and symptom inquiries.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase