prompt-agnostic-language-models/Llama-8B_single_longer

Llama 3.1 8B Instruction-Tuned Model

This model is an 8 billion parameter instruction-tuned variant from Meta's Llama 3.1 family, designed for multilingual dialogue. It leverages an optimized transformer architecture and has been fine-tuned using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.

Key Capabilities & Features

• Multilingual Support: Optimized for 8 languages including English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai, with potential for fine-tuning in others.
• Extended Context Window: Features a 128k token context length, enabling processing of longer inputs and generating more extensive responses.
• Robust Training: Pretrained on over 15 trillion tokens of publicly available online data, with a knowledge cutoff of December 2023.
• Enhanced Performance: Demonstrates improved scores over its predecessor, Llama 3 8B Instruct, across various benchmarks including MMLU (73.0 vs 65.3 CoT), HumanEval (72.6 vs 60.4 pass@1), and MATH (51.9 vs 29.1 final_em).
• Tool Use: Shows significant improvements in tool-use benchmarks like API-Bank (82.6 vs 48.3 acc) and BFCL (76.1 vs 60.3 acc).

Ideal Use Cases

• Multilingual Chatbots: Excellent for building assistant-like conversational AI systems that operate across multiple languages.
• Code Generation & Reasoning: Suitable for applications requiring strong performance in coding tasks and complex reasoning.
• Research & Commercial Applications: Intended for both research and commercial use, with a custom Llama 3.1 Community License.
• Synthetic Data Generation: Can be leveraged to improve other models through synthetic data generation and distillation.