Model Overview

This model, aifeifei799/Llama-3.1-8B-Instruct-Fei-v1-Uncensored, is an uncensored instruction-tuned version of Meta's Llama 3.1 8B. It leverages the Llama 3.1 architecture, an 8 billion parameter auto-regressive language model with an optimized transformer design, supporting a substantial 128k context length. The base Llama 3.1 models were trained on over 15 trillion tokens of publicly available online data, with a knowledge cutoff of December 2023, and fine-tuned using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF).

Key Capabilities

• Uncensored Instruction Following: Specifically modified to provide uncensored responses, offering greater flexibility for developers.
• Multilingual Support: Supports English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai, with potential for other languages through fine-tuning.
• Extended Context Window: Features a 128k token context length, enabling processing of longer inputs and generating more extensive outputs.
• Improved Performance: The Llama 3.1 8B Instruct model shows notable improvements over Llama 3 8B Instruct in benchmarks like MMLU (69.4 vs 68.5), HumanEval (72.6 vs 60.4), and MATH (51.9 vs 29.1).

Good For

• Unrestricted Dialogue Applications: Ideal for use cases requiring responses without inherent censorship or safety guardrails present in standard instruction-tuned models.
• Research and Development: Provides a valuable resource for studying the impact of uncensored models and developing custom safety policies.
• Multilingual Chatbots and Assistants: Suitable for building assistant-like chat applications across its supported languages, especially where a less restrictive output is desired.