AIbase
EN
Home
Llama2 7b Mmlu
L

Llama2 7b Mmlu

Developed by itsliupeng
Continuously trained on the MMLU dataset based on the Llama-2-7b-hf model to improve MMLU metrics while maintaining stability in other indicators
Large Language Model
Transformers
 EnglishOpen Source License:Apache-2.0#MMLU performance optimization#Multi-task text generation#Low-resource efficient inference
Downloads 120
Release Time : 10/10/2023

Model Overview

This model is an improved version of Llama-2-7b-hf, continuously trained on the mmlu_recall dataset, focusing on enhancing performance in MMLU benchmark tests while ensuring other capability metrics remain unaffected.

Model Features

MMLU performance improvement
Through continuous training on the mmlu_recall dataset, the MMLU metric reached 60.04, showing significant improvement compared to the original version
Multi-task capability retention
While improving MMLU performance, it maintains stable performance in other benchmark tests such as ARC and HellaSwag
Open-source license
Adopts the Apache-2.0 license, allowing for commercial and research use

Model Capabilities

Text generation
Knowledge Q&A
Language understanding
Reasoning ability

Use Cases

Education
Academic Q&A system
Used to answer various academic questions, especially those requiring broad knowledge
Excellent performance in MMLU benchmark tests
Research
Model performance research
Study the impact of continuous training on specific metrics
Achieved improvement in specific metrics without affecting other capabilities
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase