Model Overview

NousResearch/Meta-Llama-3.1-8B-Instruct is an 8 billion parameter instruction-tuned model from Meta's Llama 3.1 family, released on July 23, 2024. It is built on an optimized transformer architecture and fine-tuned using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. The model was pretrained on over 15 trillion tokens of publicly available online data, with a knowledge cutoff of December 2023, and features a substantial 128K context length.

Key Capabilities

• Multilingual Dialogue: Optimized for assistant-like chat in multiple languages, including English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
• Enhanced Performance: Outperforms many open-source and closed chat models on common industry benchmarks, with notable improvements in reasoning, code generation, and mathematical tasks compared to its predecessor, Llama 3 8B Instruct.
• Tool Use: Demonstrates significant advancements in tool-use benchmarks like API-Bank (82.6% accuracy) and BFCL (76.1% accuracy).
• Commercial & Research Use: Intended for a wide range of commercial and research applications, including synthetic data generation and model distillation.

When to Use This Model

• Multilingual Chatbots: Ideal for developing assistant-like applications requiring robust performance in supported languages.
• Code Generation & Math: Suitable for tasks involving code generation (HumanEval pass@1 of 72.6%) and complex mathematical problem-solving (MATH CoT of 51.9%).
• Research & Development: Valuable for researchers exploring advanced LLM capabilities, especially in multilingual contexts and tool integration.
• Applications requiring long context: Its 128K context window makes it suitable for processing and generating longer texts.