jondurbin/airoboros-l2-70b-gpt4-1.4.1
jondurbin/airoboros-l2-70b-gpt4-1.4.1 is a 69 billion parameter Llama 2 model fine-tuned by jondurbin. This model leverages a dataset generated via OpenAI's GPT-4 API, focusing on instruction-following capabilities. It is designed for general-purpose conversational AI, offering enhanced performance through its GPT-4 derived training data.
Loading preview...
Overview
jondurbin/airoboros-l2-70b-gpt4-1.4.1 is a 69 billion parameter Llama 2 model, fine-tuned using a unique dataset generated through OpenAI's GPT-4 API. This approach aims to imbue the model with advanced instruction-following and conversational abilities, building upon the robust Llama 2 architecture. The fine-tuning data was created using the airoboros dataset generation tool, which allows for the creation of specific training data types.
Key Characteristics
- Base Model: Llama 2 70B, providing a strong foundation for general language understanding and generation.
- Training Data: Utilizes a custom dataset (
jondurbin/airoboros-gpt4-1.4.1) derived from OpenAI's GPT-4 API, enhancing its instruction-following and response quality. - Context Length: Supports a context length of 32768 tokens, enabling the processing of longer inputs and generating more coherent, extended responses.
Licensing and Usage Considerations
The model's licensing is complex due to its reliance on the Llama 2 base model's custom Meta license and the use of OpenAI API-generated data. Users must comply with Meta's original license. Commercial use is advised against due to potential conflicts with OpenAI's Terms of Service regarding the use of API output for training competing models. The developer explicitly states that by using the model, users agree to indemnify them.