Dolphin 2.9 Llama 3 70b Overview

Dolphin 2.9 Llama 3 70b is a 70 billion parameter large language model, fine-tuned by Eric Hartford, Lucas Atkins, and Fernando Fernandes. It is built upon Meta's Llama-3-70b base model and maintains an 8192-token context length, consistent with its qLoRA fine-tuning. The model was trained over 2.5 days on an 8xH100 node provided by Crusoe Cloud.

Key Capabilities

Instruction Following: Designed to accurately follow user instructions.
Conversational Skills: Capable of engaging in natural and coherent dialogues.
Coding Abilities: Possesses strong coding skills, making it suitable for development tasks.
Agentic Abilities & Function Calling: Includes initial support for agentic workflows and function calling, enabling more complex interactions.
Uncensored Nature: The model is uncensored, with its dataset filtered to remove alignment and bias, resulting in high compliance to requests. Users are advised to implement their own alignment layers.

Prompt Format

The model utilizes the ChatML prompt template format for interactions.

Licensing

Dolphin 2.9 Llama 3 70b is licensed under Meta's Llama 3 Community License Agreement, permitting various uses, including commercial applications, in accordance with the license terms.