Model Overview
The andreaskoepf/llama2-13b-oasst-baseline is a 13 billion parameter language model built upon the Llama 2 architecture. This model has been fine-tuned by andreaskoepf, leveraging the Open-Assistant dataset for instruction-following capabilities. The fine-tuning process aims to enhance its performance in conversational contexts and general assistant-like interactions.
Key Features
- Llama 2 Architecture: Based on the robust Llama 2 foundation, providing strong language understanding and generation.
- Instruction-Tuned: Fine-tuned with the Open-Assistant dataset, optimizing it for following user instructions and engaging in dialogue.
- ChatML Prompt Format: Designed to work seamlessly with the ChatML prompt template, facilitating structured conversations with system, user, and assistant roles.
Usage and Prompting
This model expects input formatted according to the ChatML specification. The prompt structure includes distinct tags for system messages, user prompts, and assistant answers, ensuring clear role separation in multi-turn conversations. An example of the expected format is:
<|im_start|>system
{system message}<|im_end|>
<|im_start|>user
{user prompt}<|im_end|>
<|im_start|>assistant
{Assistant answer}<|im_end|>
Potential Use Cases
- Conversational Agents: Developing chatbots or virtual assistants that can understand and respond to user queries.
- Instruction Following: Tasks requiring the model to adhere to specific instructions provided by the user.
- Dialogue Generation: Creating interactive dialogue systems for various applications.