L

Llama 2 13b Chat Hf

Developed by meta-llama
Llama 2 is Meta's open-source 13-billion-parameter conversation-optimized large language model, aligned with human preferences using RLHF, supporting 4k context length
Downloads 147.25k
Release Time : 7/13/2023

Model Overview

A text generation model optimized for conversation scenarios, achieving safety and utility levels comparable to mainstream proprietary models

Model Features

Conversation optimization
Specifically optimized for dialogue capabilities through supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF)
Enhanced safety
Fine-tuned on TruthfulQA benchmark achieving 64.14% truthful generation rate, with harmful content generation rate below 0.01%
Efficient inference
13B parameter scale balances performance and efficiency, supporting 4k context window

Model Capabilities

Multi-turn dialogue generation
Instruction following
Knowledge Q&A
Text completion
Code generation

Use Cases

Intelligent assistant
Customer service bot
Handles user inquiries and provides accurate responses
Human evaluation shows performance close to ChatGPT
Education
Learning assistance
Explains complex concepts and generates practice questions
Scores 54.8 on MMLU comprehensive academic benchmark
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase