Overview
PericlesSavio/llama2-13b-chat-hf-finetuned is a 13 billion parameter variant from Meta's Llama 2 family of large language models. This specific model is a fine-tuned version, optimized for dialogue and chat applications, and converted for the Hugging Face Transformers format. Llama 2 models were trained on 2 trillion tokens of publicly available data, with fine-tuning data including over one million new human-annotated examples.
Key Capabilities
- Dialogue Optimization: Specifically fine-tuned using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety in chat scenarios.
- Performance: Llama-2-Chat models, including this 13B variant, have shown to outperform open-source chat models on most tested benchmarks and are competitive with some popular closed-source models like ChatGPT and PaLM in human evaluations for helpfulness and safety.
- Context Length: Features a 4k token context length, suitable for extended conversational turns.
- Commercial Use: Licensed for both commercial and research applications.
Intended Use Cases
- Assistant-like Chat: Ideal for building conversational AI agents and chatbots.
- Natural Language Generation: While fine-tuned for chat, the underlying Llama 2 architecture can be adapted for various text generation tasks.
Limitations
- English Only: Intended for use exclusively in English.
- Safety Considerations: As with all LLMs, potential for inaccurate, biased, or objectionable responses; requires safety testing and tuning for specific applications.